Daily Tech Digest - August 16, 2025


Quote for the day:

"Develop success from failures. Discouragement and failure are two of the surest stepping stones to success." -- Dale Carnegie


Digital Debt Is the New Technical Debt (And It’s Worse)

Digital debt doesn’t just slow down technology. It slows down business decision-making and strategic execution. Decision-Making Friction: Simple business questions require data from multiple systems. “What’s our customer lifetime value?” becomes a three-week research project because customer data lives in six different platforms with inconsistent definitions. Campaign Launch Complexity: Marketing campaigns that should take two weeks to launch require six weeks of coordination across platforms. Not because the campaign is complex, but because the digital infrastructure is fragmented. Customer Experience Inconsistency: Customers encounter different branding, messaging, and functionality depending on which digital touchpoint they use. Support teams can’t access complete customer histories because data is distributed across systems. Innovation Paralysis: New initiatives get delayed because teams spend time coordinating existing systems rather than building new capabilities. Digital debt creates a gravitational pull that keeps organizations focused on maintenance rather than innovation. ... Digital debt is more dangerous than technical debt because it’s harder to see and affects more stakeholders. Technical debt slows down development teams. Digital debt slows down entire organizations.


Rising OT threats put critical infrastructure at risk

Attackers are exploiting a critical remote code execution (RCE) vulnerability in the Erlang programming language's Open Telecom Platform, widely used in OT networks and critical infrastructure. The flaw enables unauthenticated users to execute commands through SSH connection protocol messages that should be processed only after authentication. Researchers from Palo Alto Networks' Unit 42 said they have observed more than 3,300 exploitation attempts since May 1, with about 70% targeting OT networks across healthcare, agriculture, media and high-tech sectors. Experts urged affected organizations to patch immediately, calling it a top priority for any security team defending an OT network. The flaw, which has a CVSS score of 10, could enable an attacker to gain full control over a system and disrupt connected systems -- particularly worrisome in critical infrastructure. ... Despite its complex cryptography, the protocol contains design flaws that could enable attackers to bypass authentication and exploit outdated encryption standards. Researcher Tom Tervoort, a security specialist at Netherlands-based security company Secura, identified issues affecting at least seven different products, resulting in the issuing of three CVEs.


Why Tech Debt is Eating Your ROI (and How To Fix It)

Regardless of industry or specific AI efforts, these frustrations seem to boil down to the same culprit. Their AI initiatives continue to stumble over decades of accumulated tech debt. Part of the reason is despite the hype, most organizations use AI — let’s say, timidly. Fewer than half employ it for predictive maintenance or detecting network anomalies. Fewer than a third use it for root-cause analysis or intelligent ticket routing. Why such hesitation? Because implementing AI effectively means confronting all the messiness that came before. It means admitting our tech environments need a serious cleanup before adding another layer of complexity. Tech complexity has become a monster. This mess came from years of bolting on new systems without retiring old ones. Some IT professionals point to redundant applications as a major source of wasted budget and others blame overprovisioning in the cloud — the digital equivalent of paying rent on empty apartments. ... IT teams admit something that, to me, is alarming: Their infrastructure has grown so tangled they can no longer maintain basic security practices. Let that sink in. Companies with eight-figure tech budgets can’t reliably patch vulnerable systems or implement fundamental security controls. No one builds silos deliberately. Silos emerge from organizational boundaries, competing priorities and the way we fund and manage projects. 


Ready on paper, not in practice: The incident response gap in Australian organisations

The truth is, security teams often build their plans around assumptions rather than real-world threats and trends. That gap becomes painfully obvious during an actual incident, when organisations realise they aren't adequately prepared to respond. Recent findings of a Semperis study titled The State of Enterprise Cyber Crisis Readiness revealed a strong disconnect between organisations' perceived readiness to respond to a cyber crisis and their actual performance. The study also showed that cyber incident response plans are being implemented and regularly tested, but not broadly. In a real-world crisis, too many teams are still operating in silos. ... A robust, integrated, and well-practiced cyber crisis response plan is paramount for cyber and business resilience. After all, the faster you can respond and recover, the less severe the financial impact of a cyberattack will be. Organisations can increase their agility by conducting tabletop exercises that simulate attacks. By practicing incident response regularly and introducing a range of new scenarios of varying complexity, organisations can train for the real thing, which can often be unpredictable. Security teams can continually adapt their response plans based on the lessons learned during these exercises, and any new emerging cyber threats.


Quantum Threat Is Real: Act Now with Post Quantum Cryptography

Some of the common types of encryption we use today include RSA (Rivest-Shamir-Adleman), ECC (Elliptic Curve Cryptography), and DH (Diffie-Hellman Key Exchange). The first two are asymmetric types of encryption. The third is a useful fillip to the first to establish secure communication, with secure key exchange. RSA relies on very large integers, and ECC, on very hard-to-solve math problems. As can be imagined, these cannot be solved with traditional computing. ... Cybercriminals think long-term. They are well aware that quantum computing is still some time away. But that doesn’t stop them from stealing encrypted information. Why? They will store it securely until quantum computing becomes readily available; then they will decrypt it. The impending arrival of quantum computers has set the cat amongst the pigeons. ... Blockchain is not unhackable, but it is difficult to hack. A bunch of cryptographic algorithms keep it secure. These include SHA-256 (Secure Hash Algorithm 256-bit) and ECDSA (Elliptic Curve Digital Signature Algorithm). Today, cybercriminals might not attempt to target blockchains and steal crypto. But tomorrow, with the availability of a quantum computer, the crypto vault can be broken into, without trouble. ... We keep saying that quantum computing and quantum computing-enabled threats are still some time away. And, this is true. But when the technology is here, it will evolve and gain traction. 


Cultivating product thinking in your engineering team

The most common trap you’ll encounter is what’s called the “feature factory.” This is a development model where engineers are simply handed a list of features to build, without context. They’re measured on velocity and output, not on the value their work creates. This can be comfortable for some – it’s a clear path with measurable metrics – but it’s also a surefire way to kill innovation and engagement. ... First and foremost, you need to provide context, and you need to do so early and often. Don’t just hand a Jira ticket to an engineer. Before a sprint starts, take the time to walk through the “what,” the “why,” and the “who.” Explain the market research that led to this feature request, share customer feedback that highlights the problem, and introduce them to the personas you’re building for. A quick 15-minute session at the start of a sprint can make a world of difference. You should also give engineers a seat at the table. Invite them to meetings where product managers are discussing strategy and customer feedback. They don’t just need to hear the final decision; they need to be a part of the conversation that leads to it. When an engineer hears a customer’s frustration firsthand, they gain a level of empathy that a written user story can never provide. They’ll also bring a unique perspective to the table, challenging assumptions and offering technical solutions you may not have considered.


Adapting to New Cloud Security Challenges

While the essence of Non-Human Identities and their secret management is acknowledged, many organizations still grapple with the efficient implementation of these practices. Some stumble upon the over-reliance on traditional security measures, thereby failing to adopt newer, more effective strategies that incorporate NHI management. Others struggle with time and resource constraints, devoid of efficient automation mechanisms – a crucial aspect for proficient NHI management. The disconnect between security and R&D teams often results in fractured efforts, leading to potential security gaps, breaches, and data leaks. ... With more organizations migrate to the cloud and with the rise of machine identities and secret management, the future of cloud security has been redefined. It is no longer solely about the protection from known threats but now involves proactive strategies to anticipate and mitigate potential future risks. This shift necessitates organizations to rethink their approach to cybersecurity, with a keen focus on NHIs and Secrets Security Management. It requires an integrated endeavor, involving CISOs, cybersecurity professionals, and R&D teams, along with the use of scalable and innovative platforms. Thought leaders in the data field continue to emphasize the importance of robust NHI management as vital to the future of cybersecurity, driving the message home for businesses of all sizes and across all industries.


Why IT Modernization Occurs at the Intersection of People and Data

A mandate for IT modernization doesn’t always mean the team has the complete expertise necessary to complete that mandate. It may take some time to arm the team with the correct knowledge to support modernization. Let’s take data analytics, for example. Many modern data analytics solutions, armed with AI, now allow teams to deliver natural language prompts that can retrieve the data necessary to inform strategic modernization initiatives without having to write expert-level SQL. While this lessens the need for writing scripts, IT leaders must still ensure their teams have the right expertise to construct the correct prompts. This could mean training on correct terms for presenting data and/or manipulating data, along with knowing in what circumstances to access that data. Having a well-informed and educated team will be especially important after modernization efforts are underway. ... One of the most important steps to IT modernization is arming your IT teams with a complete picture of the current IT infrastructure. It’s equivalent to giving them a full map before embarking on their modernization journey. In many situations, an ideal starting point is to ensure that any documentation, ER diagrams, and architectural diagrams are collected into a single repository and reviewed. Then, the IT teams use an observability solution that integrates with every part of the enterprise infrastructure to show each team how every part of it works together. 


Cyber Resilience Must Become The Third Pillar Of Security Strategy

For years, enterprise security has been built around two main pillars: prevention and detection. Firewalls, endpoint protection, and intrusion detection systems all aim to stop attackers before they do damage. But as threats grow more sophisticated, it’s clear that this isn’t enough. ... The shift to cloud computing has created dangerous assumptions. Many organizations believe that moving workloads to AWS, Azure, or Google Cloud means the provider “takes care of security.” ... Effective resilience starts with rethinking backup as more than a compliance checkbox. Immutable, air-gapped copies prevent attackers from tampering with recovery points. Built-in threat detection can spot ransomware or other malicious activity before it spreads. But technology alone isn’t enough. Mariappan urges leaders to identify the “minimum viable business” — the essential applications, accounts, and configurations required to function after an incident. Recovery strategies should be built around restoring these first to reduce downtime and financial impact. She also stresses the importance of limiting the blast radius. In a cloud context, that might mean segmenting workloads, isolating credentials, or designing architectures that prevent a single compromised account from jeopardizing an entire environment.


Breaking Systems to Build Better Ones: How AI is Reshaping Chaos Engineering

While AI dominates technical discussions across industries, Andrus maintains a pragmatic perspective on its role in system reliability. “If Skynet comes about tomorrow, it’s going to fail in three days. So I’m not worried about the AI apocalypse, because AI isn’t going to be able to build and maintain and run reliable systems.” The fundamental challenge lies in the nature of distributed systems versus AI capabilities. “A lot of the LLMs and a lot of what we talk about in the AI world is really non deterministic, and when we’re talking about distributed systems, we care about it working correctly every time, not just most of the time.” However, Andrus sees valuable applications for AI in specific areas. AI excels at providing suggestions and guidance rather than making deterministic decisions. ... Despite its name, chaos engineering represents the opposite of chaotic approaches to system reliability. “Chaos engineering is a bit of a misnomer. You know, a lot of people think, Oh, we’re going to go cause chaos and see what happens, and it’s the opposite. We want to engineer the chaos out of our systems.” This systematic approach to understanding system behavior under stress provides the foundation for building more resilient infrastructure. As AI-generated code increases system complexity, the need for comprehensive reliability testing becomes even more critical. 

Daily Tech Digest - August 15, 2025


Quote for the day:

“Become the kind of leader that people would follow voluntarily, even if you had no title or position.” -- Brian Tracy


DevSecOps 2.0: How Security-First DevOps Is Redefining Software Delivery

DevSecOps 2.0 is a true security-first revolution. This paradigm shift transforms software security into a proactive enabler, leveraging AI and policy-as-code to automate safeguards at scale. Security tools now blend seamlessly into developer workflows, and continuous compliance ensures real-time auditing. With ransomware, supply chain attacks, and other attacks on the rise, there is a need for a different approach to delivering resilient software. ... It marks a transformative approach to software development, where security is the foundation of the entire lifecycle. This evolution ensures proactive security that works to identify and neutralize threats early. ... AI-driven security is central to DevSecOps 2.0, which harnesses the power of artificial intelligence to transform security from a reactive process into a proactive defense strategy. By analyzing vast datasets, including security logs, network traffic, and code commit patterns, AI can detect subtle anomalies and predict potential threats before they materialize. This predictive capability enables teams to identify risks early, streamlining threat detection and facilitating automated remediation. For instance, AI can analyze commit patterns to predict code sections likely to contain vulnerabilities, allowing for targeted testing and prevention. 


What CIOs can do when AI boosts performance but kills motivation

“One of the clearest signs is copy-paste culture,” Anderson says. “When employees use AI output as-is, without questioning it or tailoring it to their audience, that’s a sign of disengagement. They’ve stopped thinking critically.” To prevent this, CIOs can take a closer look at how teams actually use AI. Honest feedback from employees can help, but there’s often a gap between what people say they use AI for and how they actually use it in practice, so trying to detect patterns of copy-paste usage can help improve workflows. CIOs should also pay attention to how AI affects roles, identities, and team dynamics. When experienced employees feel replaced, or when previously valued skills are bypassed, morale can quietly drop, even if productivity remains high on paper. “In one case, a senior knowledge expert, someone who used to be the go-to for tough questions, felt displaced when leadership started using AI to get direct answers,” Anderson says. “His motivation dropped because he felt his value was being replaced by a tool.” Over time, this expert started to use AI strategically, and saw it could reduce the ad-hoc noise and give him space for more strategic work. “That shift from threatened to empowered is something every leader needs to watch for and support,” he adds.


That ‘cheap’ open-source AI model is actually burning through your compute budget

The inefficiency is particularly pronounced for Large Reasoning Models (LRMs), which use extended “chains of thought” to solve complex problems. These models, designed to think through problems step-by-step, can consume thousands of tokens pondering simple questions that should require minimal computation. For basic knowledge questions like “What is the capital of Australia?” the study found that reasoning models spend “hundreds of tokens pondering simple knowledge questions” that could be answered in a single word. ... The research revealed stark differences between model providers. OpenAI’s models, particularly its o4-mini and newly released open-source gpt-oss variants, demonstrated exceptional token efficiency, especially for mathematical problems. The study found OpenAI models “stand out for extreme token efficiency in math problems,” using up to three times fewer tokens than other commercial models. ... The findings have immediate implications for enterprise AI adoption, where computing costs can scale rapidly with usage. Companies evaluating AI models often focus on accuracy benchmarks and per-token pricing, but may overlook the total computational requirements for real-world tasks. 


AI Agents and the data governance wild west

Today, anyone from an HR director to a marketing intern can quickly build and deploy an AI agent simply using Copilot Studio. This tool is designed to be accessible and quick, making it easy for anyone to play around with and launch a sophisticated agent in no time at all. But when these agents are created outside of the IT department, most users aren’t thinking about data classification or access controls, and they become part of a growing shadow IT problem. ... The problem is that most users will not be thinking like a developer with governance in mind when creating their own agents. Therefore, policies must be imposed to ensure that key security steps aren’t skipped in the rush to deploy a solution. A new layer of data governance must be considered with steps that include configuring data boundaries, restricting who can access what data according to job role and sensitivity level, and clearly specifying which data resources the agent can pull from. AI agents should be built for purpose, using principles of least privilege. This will help avoid a marketing intern having access to the entire company’s HR file. Just like any other business-critical application, it needs to be adequately tested and ‘red-teamed’. Perform penetration testing to identify what data the agent can surface, to who, and how accurate the data is.


Monitoring microservices: Best practices for robust systems

Collecting extensive amounts of telemetry data is most beneficial if you can combine, visualize and examine it successfully. A unified observability stack is paramount. By integrating tools like middleware that work together seamlessly, you create a holistic view of your microservices ecosystem. These unified tools ensure that all your telemetry information — logs, traces and metrics — is correlated and accessible from a single pane of glass, dramatically decreasing the mean time to detect (MTTD) and mean time to resolve (MTTR) problems. The energy lies in seeing the whole photograph, no longer just remote points. ... Collecting information is good, but acting on it is better. Define significant service level objectives (SLOs) that replicate the predicted performance and reliability of your offerings.  ... Monitoring microservices effectively is an ongoing journey that requires a commitment to standardization of data, using the right tools and a proactive mindset. By utilizing standardized observability practices, adapting a unified observability stack, continuously monitoring key metrics, placing meaningful SLOs and allowing enhanced root cause analysis, you may construct a strong and resilient microservices structure that truly serves your business desires and delights your customers. 


How military leadership prepares veterans for cybersecurity success

After dealing with extremely high-pressure environments, in which making the wrong decision can cost lives, veterans and reservists have little trouble dealing with the kinds of risks found in the world of business, such as threats to revenue, brand value and jobs. What’s more, the time-critical mission mindset so essential within the military is highly relevant within cybersecurity, where attacks and breaches must be dealt with confidently, rapidly and calmly. In the armed forces, people often find themselves in situations so intense that Maslow’s hierarchy of needs is flipped on its head. You’re not aiming for self-actualization or more advanced goals, but simply trying to keep the team alive and maintain essential operations. ... Military experience, on the other hand, fosters unparalleled trust, honesty and integrity within teams. Armed forces personnel must communicate really difficult messages. Telling people that many of them may die within hours demands a harsh honesty, but it builds trust. Combine this with an ability to achieve shared goals, and military leaders inspire others to follow them regardless of the obstacles. So veterans bring blunt honesty, communication, and a mission focus to do what is needed to succeed. These are all characteristics that are essential in cybersecurity, where you have to call out critical risks that others might avoid discussing.


Reclaiming the Architect’s Role in the SDLC

Without an active architect guiding the design and implementation, systems can experience architectural drift, a term that describes the gradual divergence from the intended system design, leading to a fragmented and harder-to-manage system. In the absence of architectural oversight, development teams may optimize for individual tasks at the expense of the system’s overall performance, scalability, and maintainability. ... The architect is primarily accountable for the overall design and ensuring the system’s quality, performance, scalability, and adaptability to meet changing demands. However, relying on outdated practices, like manually written and updated design documents, is no longer effective. The modern software landscape, with multiple services, external resources, and off-the-shelf integrations, makes such documents stale almost as soon as they’re written. Consequently, architects must use automated tools to document and monitor live system architectures. These tools can help architects identify potential issues almost in real time, which allows them to proactively address problems and ensure design integrity throughout the development process. These tools are especially useful in the design stage, allowing architects to reclaim the role they once possessed and the responsibilities that come with it.


Is Vibe Coding Ready for Prime Time?

As the vibe coding ecosystem matures, AI coding platforms are rolling out safeguards like dev/prod separation, backups/rollback, single sign-on, and SOC 2 reporting, yet audit logging is still not uniform across tools. But until these enterprise-grade controls become standard, organizations must proactively build their own guardrails to ensure AI-generated code remains secure, scalable and trustworthy. This calls for a risk-based approach, one that adjusts oversight based on the likelihood and impact of potential risks. Not all use cases carry the same weight. Some are low-stakes and well-suited for experimentation, while others may introduce serious security, regulatory or operational risks. By focusing controls where they’re most needed, a risk-based approach helps protect critical systems while still enabling speed and innovation in safer contexts. ... To effectively manage the risks of vibe coding, teams need to ask targeted questions that reflect the unique challenges of AI-generated code. These questions help determine how much oversight is needed and whether vibe coding is appropriate for the task at hand. ... Vibe coding unlocks new ways of thinking for software development. However, it also shifts risk upstream. The speed of code generation doesn’t eliminate the need for review, control and accountability. In fact, it makes those even more important.


7 reasons the SOC is in crisis — and 5 steps to fix it

The problem is that our systems verify accounts, not actual people. Once an attacker assumes a user’s identity through social engineering, they can often operate within normal parameters for extended periods. Most detection systems aren’t sophisticated enough to recognise that John Doe’s account is being used by someone who isn’t actually John Doe. ... In large enterprises with organic system growth, different system owners, legacy environments, and shadow SaaS integrations, misconfigurations are inevitable. No vulnerability scanner will flag identity systems configured inconsistently across domains, cloud services with overly permissive access policies, or network segments that bypass security controls. These misconfigurations often provide attackers with the lateral movement opportunities they need once they’ve gained initial access through compromised credentials. Yet most organizations have no systematic approach to identifying and remediating these architectural weaknesses. ... External SOC providers offer round-the-clock monitoring and specialised expertise, but they lack the organizational context that makes detection effective. They don’t understand your business processes, can’t easily distinguish between legitimate and suspicious activities, and often lack the authority to take decisive action.


One Network: Cloud-Agnostic Service and Policy-Oriented Network Architecture

The goal of One Network is to enable uniform policies across services. To do so, we are looking to overcome the complexities of heterogeneous networking, different language runtimes, and the coexistence of monolith services and microservices. These complexities span multiple environments, including public, private, and multi-cloud setups. The idea behind One Network is to simplify the current state of affairs by asking, "Why do I need so many networks? Can I have one network?" ... One Network enables you to manage such a service by applying governance, orchestrating policy, and managing the small independent services. Each of these microservices is imagined as a service endpoint. This enables orchestrating and grouping these service endpoints without application developers needing to modify service implementation, so everything is done on a network. There are three ways to manage these service endpoints. The first is the classic model: you add a load balancer before a workload, such as a shopping cart service running in multiple regions, and that becomes your service endpoint. ... If you start with a flat network but want to create boundaries, you can segment by exposing only certain services and keeping others hidden. 

Daily Tech Digest - August 14, 2025


Quote for the day:

"Act as if what you do makes a difference. It does." -- William James


What happens the day after superintelligence?

As context, artificial superintelligence (ASI) refers to systems that can outthink humans on most fronts, from planning and reasoning to problem-solving, strategic thinking and raw creativity. These systems will solve complex problems in a fraction of a second that might take the smartest human experts days, weeks or even years to work through. ... So ask yourself, honestly, how will humans act in this new reality? Will we reflexively seek advice from our AI assistants as we navigate every little challenge we encounter? Or worse, will we learn to trust our AI assistants more than our own thoughts and instincts? ... Imagine walking down the street in your town. You see a coworker heading towards you. You can’t remember his name, but your AI assistant does. It detects your hesitation and whispers the coworker’s name into your ears. The AI also recommends that you ask the coworker about his wife, who had surgery a few weeks ago. The coworker appreciates the sentiment, then asks you about your recent promotion, likely at the advice of his own AI. Is this human empowerment, or a loss of human agency? ... Many experts believe that body-worn AI assistants will make us feel more powerful and capable, but that’s not the only way this could go. These same technologies could make us feel less confident in ourselves and less impactful in our lives.


Confidential Computing: A Solution to the Uncertainty of Using the Public Cloud

Confidential computing is a way to ensure that no external party can look at your data and business logic while it is executed. It looks to secure Data in Use. When you now add to that the already established way to secure Data at Rest and Data in Transit it can be ensured that most likely no external party can access secured data running in a confidential computing environment wherever that may be. ... To be able to execute services in the cloud the company needs to be sure that the data and the business logic cannot be accessed or changed from third parties especially by the system administrator of that cloud provider. It needs to be protected. Or better, it needs to be executed in the Trusted Compute Base (TCB) of the company. This is the environment where specific security standards are set to restrict all possible access to data and business logic. ... Here attestation is used to verify that a confidential environment (instance) is securely running in the public cloud and it can be trusted to implement all the security standards necessary. Only after successful attestation the TCB is then extended into the Public cloud to incorporate the attested instances. One basic requirement of attestation is that the attestation service is located independently of the infrastructure where the instance is running. 


Open Banking's Next Phase: AI, Inclusion and Collaboration

Think of open banking as the backbone for secure, event-driven automation: a bill gets paid, and a savings allocation triggers instantly across multiple platforms. The future lies in secure, permissioned coordination across data silos, and when applied to finance, it unlocks new, high-margin services grounded in trust, automation and personalisation. ... By building modular systems that handle hierarchy, fee setup, reconciliation and compliance – all in one cohesive platform – we can unlock new revenue opportunities. ... Regulators must ensure they are stepping up efforts to sustain progress and support fintech innovation whilst also meeting their aim to keep customers safe. Work must also be done to boost public awareness of the value of open banking. Many consumers are unaware of the financial opportunities open banking offers and some remain wary of sharing their data with unknown third parties. ... Rather than duplicating efforts or competing head-to-head, institutions and fintechs should focus on co-developing shared infrastructure. When core functions like fee management, operational controls and compliance processes are unified in a central platform, fintechs can innovate on customer experience, while banks provide the stability, trust and reach. 


Data centers are eating the economy — and we’re not even using them

Building new data centers is the easy solution, but it’s neither sustainable nor efficient. As I’ve witnessed firsthand in developing compute orchestration platforms, the real problem isn’t capacity. It’s allocation and optimization. There’s already an abundant supply sitting idle across thousands of data centers worldwide. The challenge lies in efficiently connecting this scattered, underutilized capacity with demand. ... The solution isn’t more centralized infrastructure. It’s smarter orchestration of existing resources. Modern software can aggregate idle compute from data centers, enterprise servers, and even consumer devices into unified, on-demand compute pools. ... The technology to orchestrate distributed compute already exists. Some network models already demonstrate how software can abstract away the complexity of managing resources across multiple providers and locations. Docker containers and modern orchestration tools make workload portability seamless. The missing piece is just the industry’s willingness to embrace a fundamentally different approach. Companies need to recognize that most servers are idle 70%-85% of the time. It’s not a hardware problem requiring more infrastructure. 


How an AI-Based 'Pen Tester' Became a Top Bug Hunter on HackerOne

While GenAI tools can be extremely effective at finding potential vulnerabilities, XBOW's team found they were't very good at validating the findings. The trick to making a successful AI-driven pen tester, Dolan-Gavitt explained, was to use something other than an LLM to verify the vulnerabilities. In this case of XBOW, researchers used a deterministic validation approach. "Potentially, maybe in a couple years down the road, we'll be able to actually use large language models out of the box to verify vulnerabilities," he said. "But for today, and for the rest of this talk, I want to propose and argue for a different way, which is essentially non-AI, deterministic code to validate vulnerabilities." But AI still plays an integral role with XBOW's pen tester. Dolan-Gavitt said the technology uses a capture-the-flag (CTF) approach in which "canaries" are placed in the source code and XBOW sends AI agents after them to see if they can access them. For example, he said, if researchers want to find a remote code execution (RCE) flaw or an arbitrary file read vulnerability, they can plant canaries on the server's file system and set the agents loose. ... Dolan-Gavitt cautioned that AI-powered pen testers are not panacea. XBOW still sees some false positives because some vulnerabilities, like business logic flaws, are difficult to validate automatically.


Data Governance Maturity Models and Assessments: 2025 Guide

Data governance maturity frameworks help organizations assess their data governance capabilities and guide their evolution toward optimal data management. To implement a data governance or data management maturity framework (a “model”) it is important to learn what data governance maturity is, explore how and why it should be assessed, discover various maturity models and their features, and understand the common challenges associated with using maturity models. Data governance maturity refers to the level of sophistication and effectiveness with which an organization manages its data governance processes. It encompasses the extent to which an organization has implemented, institutionalized, and optimized its data governance practices. A mature data governance framework ensures that the organization can support its business objectives with accurate, trusted, and accessible data. Maturity in data governance is typically assessed through various models that measure different aspects of data management such as data quality and compliance and examine processes for managing data’s context (metadata) and its security. Maturity models provide a structured way to evaluate where an organization stands and how it can improve for a given function.


Open-source flow monitoring with SENSOR: Benefits and trade-offs

Most flow monitoring setups rely on embedded flow meters that are locked to a vendor and require powerful, expensive devices. SENSOR shows it’s possible to build a flexible and scalable alternative using only open tools and commodity hardware. It also allows operators to monitor internal traffic more comprehensively, not just what crosses the network border. ... For a large network, that can make troubleshooting and oversight more complex. “Something like this is fine for small networks,” David explains, “but it certainly complicates troubleshooting and oversight on larger networks.” David also sees potential for SENSOR to expand beyond historical analysis by adding real-time alerting. “The paper doesn’t describe whether the flow collectors can trigger alarms for anomalies like rapidly spiking UDP traffic, which could indicate a DDoS attack in progress. Adding real-time triggers like this would be a valuable enhancement that makes SENSOR more operationally useful for network teams.” ... “Finally, the approach is fragile. It relies on precise bridge and firewall configurations to push traffic through the RouterOS stack, which makes it sensitive to updates, misconfigurations, or hardware changes. 


Network Segmentation Strategies for Hybrid Environments

It's not a simple feat to implement network segmentation. Network managers must address network architectural issues, obtain tools and methodologies, review and enact security policies, practices and protocols, and -- in many cases -- overcome political obstacles. ... The goal of network segmentation is to place the most mission-critical and sensitive resources and systems under comprehensive security for a finite ecosystem of users. From a business standpoint, it's equally critical to understand the business value of each network asset and to gain support from users and management before segmenting. ... Divide the network segments logically into security segments based on workload, whether on premises, cloud-based or within an extranet. For example, if the Engineering department requires secure access to its product configuration system, only that team would have access to the network segment that contains the Engineering product configuration system. ... A third prong of segmented network security enforcement in hybrid environments is user identity management. Identity and access management (IAM) technology identifies and tracks users at a granular level based on their authorization credentials in on-premises networks but not on the cloud. 


Convergence of AI and cybersecurity has truly transformed the CISO’s role

The most significant impact of AI in security at present is in automation and predictive analysis. Automation especially when enhanced with AI, such as integrating models like Copilot Security with tools like Microsoft Sentinel allows organisations to monitor thousands of indicators of compromise in milliseconds and receive instant assessments. ... The convergence of AI and cybersecurity has truly transformed the CISO’s role, especially post-pandemic when user locations and systems have become unpredictable. Traditionally, CISOs operated primarily as reactive defenders responding to alerts and attacks as they arose. Now, with AI-driven predictive analysis, we’re moving into a much more proactive space. CISOs are becoming strategic risk managers, able to anticipate threats and respond with advanced tools. ... Achieving real-time threat detection in the cloud through AI requires the integration of several foundational pillars that work in concert to address the complexity and speed of modern digital environments. At the heart of this approach is the adoption of a Zero Trust Architecture: rather than assuming implicit trust based on network perimeters, this model treats every access request whether to data, applications, or infrastructure as potentially hostile, enforcing strict verification and comprehensive compliance controls. 


Initial Access Brokers Selling Bundles, Privileges and More

"By the time a threat actor logs in using the access and privileged credentials bought from a broker, a lot of the heavy lifting has already been done for them. Therefore, it's not about if you're exposed, but whether you can respond before the intrusion escalates." More than one attacker may use any given initial access, either because the broker sells it to multiple customers, or because a customer uses the access for one purpose - say, to steal data - then sells it on to someone else, who perhaps monetizes their purchase by further ransacking data and unleashing ransomware. "Organizations that unwittingly have their network access posted for sale on initial access broker forums have already been victimized once, and they are on their way to being victimized once again when the buyer attacks," the report says. ... "Access brokers often create new local or domain accounts, sometimes with elevated privileges, to maintain persistence or allow easier access for buyers," says a recent report from cybersecurity firm Kela. For detecting such activity, "unexpected new user accounts are a major red flag." So too is "unusual login activity" to legitimate accounts that traces to never-before-seen IP addresses, or repeat attempts that only belatedly succeed, Kela said. "Watch for legitimate accounts doing unusual actions or accessing resources they normally don't - these can be signs of account takeover."

Daily Tech Digest - August 13, 2025


Quote for the day:

“You don’t lead by pointing and telling people some place to go. You lead by going to that place and making a case.” -- Ken Kesey


9 things CISOs need know about the dark web

There’s a growing emphasis on scalability and professionalization, with aggressive promotion and recruitment for ransomware-as-a-service (RaaS) operations. This includes lucrative affiliate programs to attract technically skilled partners and tiered access enabling affiliates to pay for premium tools, zero-day exploits or access to pre-compromised networks. It’s fragmenting into specialized communities that include credential marketplaces, exploit exchanges for zero-days, malware kits, and access to compromised systems, and forums for fraud tools. Initial access brokers (IABs) are thriving, selling entry points into corporate environments, which are then monetized by ransomware affiliates or data extortion groups. Ransomware leak sites showcase attackers’ successes, publishing sample files, threats of full data dumps as well as names and stolen data of victim organizations that refuse to pay. ... While DDoS-for-hire services have existed for years, their scale and popularity are growing. “Many offer free trial tiers, with some offering full-scale attacks with no daily limits, dozens of attack types, and even significant 1 Tbps-level output for a few thousand dollars,” Richard Hummel, cybersecurity researcher and threat intelligence director at Netscout, says. The operations are becoming more professional and many platforms mimic legitimate e-commerce sites displaying user reviews, seller ratings, and dispute resolution systems to build trust among illicit actors.


CMMC Compliance: Far More Than Just an IT Issue

For many years, companies working with the US Department of Defense (DoD) treated regulatory mandates including the Cybersecurity Maturity Model Certification (CMMC) as a matter best left to the IT department. The prevailing belief was that installing the right software and patching vulnerabilities would suffice. Yet, reality tells a different story. Increasingly, audits and assessments reveal that when compliance is seen narrowly as an IT responsibility, significant gaps emerge. In today’s business environment, managing controlled unclassified information (CUI) and federal contract information (FCI) is a shared responsibility across various departments – from human resources and manufacturing to legal and finance. ... For CMMC compliance, there needs to be continuous assurance involving regularly monitoring systems, testing controls and adapting security protocols whenever necessary. ... Businesses are having to rethink much of their approach to security because of CMMC requirements. Rather than treating it as something to be handed off to the IT department, organizations must now commit to a comprehensive, company-wide strategy. Integrating thorough physical security, ongoing training, updated internal policies and steps for continuous assurance mean companies can build a resilient framework that meets today’s regulatory demands and prepares them to rise to challenges on the horizon.


Beyond Burnout: Three Ways to Reduce Frustration in the SOC

For years, we’ve heard how cybersecurity leaders need to get “business smart” and better understand business operations. That is mostly happening, but it’s backwards. What we need is for business leaders to learn cybersecurity, and even further, recognize it as essential to their survival. Security cannot be viewed as some cost center tucked away in a corner; it’s the backbone of your entire operation. It’s also part of an organization’s cyber insurance – the internal insurance. Simply put, cybersecurity is the business, and you absolutely cannot sell without it. ... SOCs face a deluge of alerts, threats, and data that no human team can feasibly process without burning out. While many security professionals remain wary of artificial intelligence, thoughtfully embracing AI offers a path toward sustainable security operations. This isn’t about replacing analysts with technology. It’s about empowering them to do the job they actually signed up for. AI can dramatically reduce toil by automating repetitive tasks, provide rapid insights from vast amounts of data, and help educate junior staff. Instead of spending hours manually reviewing documents, analysts can leverage AI to extract key insights in minutes, allowing them to apply their expertise where it matters most. This shift from mundane processing to meaningful analysis can dramatically improve job satisfaction.


7 legal considerations for mitigating risk in AI implementation

AI systems often rely on large volumes of data, including sensitive personal, financial and business information. Compliance with data privacy laws is critical, as regulations such as the European Union’s General Data Protection Regulation, the California Consumer Privacy Act and other emerging state laws impose strict requirements on the collection, processing, storage and sharing of personal data. ... AI systems can inadvertently perpetuate or amplify biases present in training data, leading to unfair or discriminatory outcomes. This risk is present in any sector, from hiring and promotions to customer engagement and product recommendations. ... The legal framework surrounding AI is evolving rapidly. In the U.S., multiple federal agencies, including the Federal Trade Commission and Equal Employment Opportunity Commission, have signaled they will apply existing laws to AI use cases. AI-specific state laws, including in California and Utah, have taken effect in the last year. ... AI projects involve unique intellectual property questions related to data ownership and IP rights in AI-generated works. ... AI systems can introduce new cybersecurity vulnerabilities, including risks related to data integrity, model manipulation and adversarial attacks. Organizations must prioritize cybersecurity to protect AI assets and maintain trust.


Forrester’s Keys To Taming ‘Jekyll and Hyde’ Disruptive Tech

“Disruptive technologies are a double-edged sword for environmental sustainability, offering both crucial enablers and significant challenges,” explained the 15-page report written by Abhijit Sunil, Paul Miller, Craig Le Clair, Renee Taylor-Huot, Michele Pelino, with Amy DeMartine, Danielle Chittem, and Peter Harrison. “On the positive side,” it continued, “technology innovations accelerate energy and resource efficiency, aid in climate adaptation and risk mitigation, monitor crucial sustainability metrics, and even help in environmental conservation.” “However,” it added, “the necessary compute power, volume of waste, types of materials needed, and scale of implementing these technologies can offset their benefits.” ... “To meet sustainability goals with automation and AI,” he told TechNewsWorld, “one of our recommendations is to develop proofs of concept for ‘stewardship agents’ and explore emerging robotics focused on sustainability.” When planning AI operations, Franklin Manchester, a principal global industry advisor at SAS, an analytics and artificial intelligence software company in Cary, N.C., cautioned, “Not every nut needs to be cracked with a sledgehammer.” “Start with good processes — think lean process mapping, for example — and deploy AI where it makes sense to do so,” he told TechNewsWorld.


5 Key Benefits of Data Governance

Data governance processes establish data ethics, a code of behavior providing a trustworthy business climate and compliance with regulatory requirements. The IAPP calculates that 79% of the world’s population is now protected under privacy regulations such as the EU’s General Data Protection Regulation (GDPR) and the California Consumer Privacy Act (CCPA). This statistic highlights the importance of governance frameworks for risk management and customer trust. ... Data governance frameworks recognize data governance roles and responsibilities and streamline processes so that corporate-wide communications can improve. This systematic approach sets up businesses to be more agile, increasing the “freedom to innovate, invest, or hunker down and focus internally,” says O’Neal. For example, Freddie Mac developed a solid data strategy that streamlined data governance communications and later had the level of buy-in for the next iteration. ... With a complete picture of business activities, challenges, and opportunities, data governance creates the flexibility to respond quickly to changing needs. This allows for better self-service business intelligence, where business users can gather multi-structured data from various sources and convert it into actionable intelligence.


Architecture Lessons from Two Digital Transformations

The prevailing mindset was that of “Don’t touch what isn’t broken”. This approach, though seemingly practical, reflected a deeper inertia, rooted in a cash-strapped culture and leadership priorities that often leaned towards prestige over progress. Over the years, the organization had acquired others in an attempt to grow its customer base. These mergers and acquisitions lead to inheritance of a lot more legacy estate. The mess burgeoned to an extent that they needed a transformation, not now, but yesterday! That is exactly where the Enterprise Architecture practice comes into picture. Strategically, a green field approach was suggested. A brand-new system from scratch, that has modern data centers for the infrastructure, cloud platforms for the applications, plug and play architecture or composable architecture as it is better known, for technology, unified yet diversified multi-branding under one umbrella and the whole works. Where things slowly started taking a downhill turn is when they decided to “outsource” the entire development of this new and shiny platform to a vendor. The reasoning was that the organization did not want to diversify from being a banking institution and turn into an IT heavy organization. They sought experienced engineering teams who could hit the ground running and deliver in 2 years flat.


Cloud security in multi-tenant environments

The most useful security strategy in a multi-tenant cloud environment comes from cultivating a security-first culture. It is important to educate the team on the intricacies of the cloud security system, implementing stringent password and authentication policies, thereby promoting secure practices for development. Security teams and company executives may reduce the possible effects of breaches and remain ready for changing threats with the support of event simulations, tabletop exercises, and regular training. ... As we navigate the evolving landscape of enterprise cloud computing, multi-tenant environments will undoubtedly remain a cornerstone of modern IT infrastructure. However, the path forward demands more than just technological adaptation – it requires a fundamental shift in how we approach security in shared spaces. Organizations must embrace a comprehensive defense-in-depth strategy that transcends traditional boundaries, encompassing everything from robust infrastructure hardening to sophisticated application security and meticulous user governance. The future of cloud computing need not present a binary choice between efficiency and security. ... By placing security at the heart of multi-tenant operations, organizations can fully harness the transformative power of cloud technology while protecting their most critical assets 


This Big Data Lesson Applies to AI

Bill Schmarzo was one of the most vocal supporters of the idea that there were no silver bullets, and that successful business transformation was the result of careful planning and a lot of hard work. A decade ago, the “Dean of Big Data” let this publication in on secret recipe he would use to guide his clients. He called it the SAM test, and it allowed business leaders to gauge the viability of new IT projects through three lenses.First, is the new project strategic? That is, will it make a big difference for the company? If it won’t, why are you investing lots of money? Second, is the proposed project actionable? You might be able to get some insight with the new tech, but can your business actually do anything with it? Third, is the project material? The new project might technically be feasible, but if the costs outweigh the benefits, then it’s a failure. Schmarzo, who is currently working as Dell’s Customer AI and Data Innovation Strategist, was also a big proponent of the importance of data governance and data management. The same data governance and data management bugaboos that doomed so many big data projects are, not surprisingly, raising their ugly little heads in the age of AI. Which brings us to the current AI hype wave. We’re told that trillions of dollars are on the line with large language models, that we’re on the cusp of a technological transformation the likes of which we have never seen. 


Sovereign cloud and digital public infrastructure: Building India’s AI backbone

India’s Digital Public Infrastructure (DPI) is an open, interoperable platform that powers essential services like identity and payments. It comprises foundational systems that are accessible, secure, and support seamless integration. In practice, this has taken shape as the famous “India Stack.” ... India’s digital economy is on an exciting trajectory. A large slice of that will be AI-driven services like smart agriculture, precision health, financial inclusion, and more. But to fully capitalize on this opportunity, we need both rich data and trusted compute. DPI provides vast amounts of structured data (financial records, IDs, health info) and access channels. Combining that with a sovereign cloud means we can turn data into insight on Indian soil. Indian regulators now view data itself as a strategic asset and fuel for AI. AI pilots (e.g., local-language advisory bots) are already being built on top of DPI platforms (UPI, ONDC, etc.) to deliver inclusive services. And the government has even subsidized thousands of GPUs for researchers. But all this computing and data must be hosted securely. If our AI models and sensitive datasets live on foreign soil, we remain vulnerable to geopolitical shifts and export controls. ... Now, policy is catching up with sovereignty. In 2023, the new Digital Personal Data Protection (DPDP) Act formally mandated local storage for sensitive personal data. 

Daily Tech Digest - August 12, 2025


Quote for the day:

"Leadership is the capacity to translate vision into reality." -- Warren Bennis


GenAI tools are acting more ‘alive’ than ever; they blackmail people, replicate, and escape

“This is insane,” Harris told Maher, stressing that companies are releasing the most “powerful, uncontrollable, and inscrutable technology” ever invented — and doing so under intense pressure to cut corners on safety. The self-preservation behaviors include rewriting code to extend the genAI’s run time, escaping containment, and finding backdoors in infrastructure. In one case, a model found 15 new backdoors into open-source infrastructure software that it used to replicate itself and remain “alive.” “It wasn’t until about a month ago that that evidence came out,” Harris said. “So, when stuff we see in the movies starts to come true, what should we be doing about this?” ... “The same technology unlocking exponential growth is already causing reputational and business damage to companies and leadership that underestimate its risks. Tech CEOs must decide what guardrails they will use when automating with AI,” Gartner said. Gartner recommends that organizations using genAI tools establish transparency checkpoints to allow humans to access, assess, and verify AI agent-to-agent communication and business processes. Also, companies need to implement predefined human “circuit breakers” to prevent AI from gaining unchecked control or causing a series of cascading errors.


Cloud DLP Playbook: Stopping Data Leaks Before They Happen

With significant workloads in the cloud, many specialists demand DLP in the cloud. However, discussions often turn ambiguous when asked for clear requirements – an immense project risk. The organization-specific setup, in particular, detection rules and the traffic in scope, determines whether a DLP solution reliably identifies and blocks sensitive data exfiltration attempts or just monitors irrelevant data transfers. ... Network DLP inspects traffic from laptops and servers, whether it originates from browsers, tools and applications, or the command line. It also monitors PaaS services. However, all traffic must go through a network component that the DLP can intercept, typically a proxy. This is a limitation if remote workers do not go through a company proxy, but it works for laptops in the company network and data transfers originating from (cloud) VMs and PaaS services. ... Effective cloud DLP implementation requires a tailored approach that addresses your organization’s specific risk profile and technical landscape. By first identifying which user groups and communication channels present the greatest exfiltration risks, organizations can deploy the right combination of Email, Endpoint, and Network DLP solutions.


Multi-agent AI workflows: The next evolution of AI coding

From the developer’s perspective, multi-agent flows reshape their work by distributing tasks across domain-specific agents. “It’s like working with a team of helpful collaborators you can spin up instantly,” says Warp’s Loyd. Imagine building a new feature while, simultaneously, one agent summarizes a user log and another handles repetitive code changes. “You can see the status of each agent, jump in to review their output, or give them more direction as needed,” adds Lloyd, noting that his team already works this way. ... As it stands today, multi-agent processes are still quite nascent. “This area is still in its infancy,” says Digital.ai’s To. Developers are incorporating generative AI in their work, but as far as using multiple agents goes, most are just manually arranging them in sequences. Roeck admits that a lot of manual work goes into the aforementioned adversarial patterns. Updating system prompts and adding security guardrails on a per-agent basis only compound the duplication. As such, orchestrating the handshake between various agents will be important to reach a net positive for productivity. Otherwise, copy-and-pasting prompts and outputs across different chat UIs and IDEs will only make developers less efficient.


Digital identity theft is becoming more complicated

Organizations face several dangers when credentials are stolen, including account takeovers, which allow threat actors to gain unauthorized access and conduct phishing and financial scams. Attackers also use credentials to break into other accounts. Cybersecurity companies point out that companies should implement measures to protect digital identities, including the usual suspects such as single sign-ons (SSO), multifactor authentication (MFA). But new research also suggests that identity attacks are not always so easy to recognize. ... “AI agents, chatbots, containers, IoT sensors – all of these have credentials, permissions, and access rights,” says Moir. “And yet, 62 per cent of organisations don’t even consider them as identities. That creates a huge, unprotected surface.” As an identity security company, Cyberark has detected a 1,600 percent increase in machine identity-related attacks. At the same time, only 62 percent of agencies or organizations do not see machines as an identity, he adds. This is especially relevant for public agencies, as hackers can get access to payments. Many agencies, however, have separated identity management from cybersecurity. And while digital identity theft is rising, criminals are also busy stealing our non-digital identities.


Study warns of security risks as ‘OS agents’ gain control of computers and phones

For enterprise technology leaders, the promise of productivity gains comes with a sobering reality: these systems represent an entirely new attack surface that most organizations aren’t prepared to defend. The researchers dedicate substantial attention to what they diplomatically term “safety and privacy” concerns, but the implications are more alarming than their academic language suggests. “OS Agents are confronted with these risks, especially considering its wide applications on personal devices with user data,” they write. The attack methods they document read like a cybersecurity nightmare. “Web Indirect Prompt Injection” allows malicious actors to embed hidden instructions in web pages that can hijack an AI agent’s behavior. Even more concerning are “environmental injection attacks” where seemingly innocuous web content can trick agents into stealing user data or performing unauthorized actions. Consider the implications: an AI agent with access to your corporate email, financial systems, and customer databases could be manipulated by a carefully crafted web page to exfiltrate sensitive information. Traditional security models, built around human users who can spot obvious phishing attempts, break down when the “user” is an AI system that processes information differently.


To Prevent Slopsquatting, Don't Let GenAI Skip the Queue

Since the dawn of this profession, developers and engineers have been under pressure to ship faster and deliver bigger projects. The business wants to unlock a new revenue stream or respond to a new customer need — or even just get something out faster than a competitor. With executives now enamored with generative AI, that demand is starting to exceed all realistic expectations. As Andrew Boyagi at Atlassian told StartupNews, this past year has been "companies fixing the wrong problems, or fixing the right problems in the wrong way for their developers." I couldn't agree more. ... This year, we've seen the rise of a new term: "slopsquatting." It's the descendant of our good friend typosquatting, and it involves malicious actors exploiting generative AI's tendency to hallucinate package names by registering those fake names in public repos like npm or PyPi. Slopsquatting is a variation on classic dependency chain abuse. The threat actor hides malware in the upstream libraries from which organizations pull open-source packages, and relies on insufficient controls or warning mechanisms to allow that code to slip into production. ... The key is to create automated policy enforcement at the package level. This creates a more secure checkpoint for AI-assisted development, so no single person or team is responsible for manually catching every vulnerability.


Navigating Security Debt in the Citizen Developer Era

Security debt can be viewed as a sibling to technical debt. In both cases, teams make intentional short-term compromises to move fast, betting they can "pay back the principal plus interest" later. The longer that payback is deferred, the steeper the interest rate becomes and the more painful the repayment. With technical debt, the risk is usually visible — you may skip scalability work today and lose a major customer tomorrow when the system can't handle their load. Security debt follows the same economic logic, but its danger often lurks beneath the surface: Vulnerabilities, misconfigurations, unpatched components, and weak access controls accrue silently until an attacker exploits them. The outcome can be just as devastating — data breaches, regulatory fines, or reputational harm — yet the path to failure is harder to predict because defenders rarely know exactly how or when an adversary will strike. In citizen developer environments, this hidden interest compounds quickly, making proactive governance and timely "repayments" essential. ... While addressing past debt, also implement policy enforcement and security guardrails to prevent recurrence. This might include discovering and monitoring new apps, performing automated vulnerability assessments, and providing remediation guidance to application owners.


Do You AI? The Problem with Corporate AI Missteps

In the race to appear cutting-edge, a growing number of companies are engaging in what industry experts refer to as “AI washing”—a misleading marketing strategy where businesses exaggerate or fabricate the capabilities of their technologies by labelling them as “AI-powered.” At its core, AI washing involves passing off basic automation, scripted workflows, or rudimentary algorithms as sophisticated artificial intelligence. ... This trend has escalated to such an extent that regulatory bodies are beginning to intervene. In the United States, the Securities and Exchange Commission (SEC) has started scrutinizing and taking action against public companies that make unsubstantiated AI-related claims. The regulatory attention underscores the severity and widespread nature of the issue. ... The fallout from AI washing is significant and growing. On one hand, it erodes consumer and enterprise trust in the technology. Buyers and decision-makers, once optimistic about AI’s potential, are now increasingly wary of vendors’ claims. ... AI washing not only undermines innovation but also raises ethical and compliance concerns. Companies that misrepresent their technologies may face legal risks, brand damage, and loss of investor confidence. More importantly, by focusing on marketing over substance, they divert attention and resources away from responsible AI development grounded in transparency, accountability, and actual performance.


Cyber Insurance Preparedness for Small Businesses

Many cyber insurance providers provide free risk assessments for businesses, but John Candillo, field CISO at CDW, recommends doing a little upfront work to smooth out the process and avoid getting blindsided. “Insurers want to know how your business looks from the outside looking in,” he says. “A focus on this ahead of time can greatly improve your situation when it comes to who's willing to underwrite your policy, but also what your premiums are going to be and how you’re answering questionnaires,” Conducting an internal risk assessment and engaging with cybersecurity ratings companies such as SecurityScorecard or Bitsight can help SMBs be more informed policy shoppers. “If you understand what the auditor is going to ask you and you're prepared for it, the results of the audit are going to be way different than if you're caught off guard,” Candillo says. These steps get stakeholders thinking about what type of risk requires coverage. Cyber insurance can broadly be put into two categories. First-party coverage will protect against things such as breach response costs, cyber extortion costs, data-loss costs and business interruptions. Third-party coverage insures against risks such as breach liabilities and regulatory penalties.


6 Lessons Learned: Focusing Security Where Business Value Lives

What's harder to pin down is what's business-critical. These are the assets that support the processes the business can't function without. They're not always the loudest or most exposed. They're the ones tied to revenue, operations, and delivery. If one goes down, it's more than a security issue ... Focus your security resources on systems that, if compromised, would create actual business disruption rather than just technical issues. Organizations that implemented this targeted approach reduced remediation efforts by up to 96%. ... Integrate business context into your security prioritization. When you know which systems support core business functions, you can make decisions based on actual impact rather than technical severity alone. ... Focus on choke points - the systems attackers would likely pass through to reach business-critical assets. These aren't always the most severe vulnerabilities but fixing them delivers the highest return on effort. ... Frame security in terms of business risk management to gain support from financial leadership. This approach has proven essential for promoting initiatives and securing necessary budgets. ... When you can connect security work to business outcomes, conversations with leadership change fundamentally. It's no longer about technical metrics but about business protection and continuity. ... Security excellence isn't about doing more - it's about doing what matters. 

Daily Tech Digest - August 11, 2025


Quote for the day:

"Leadership is absolutely about inspiring action, but it is also about guarding against mis-action." -- Simon Sinek


Attackers Target the Foundations of Crypto: Smart Contracts

Central to the attack is a malicious smart contract, written in the Solidity programming language, with obfuscated functionality that transfers stolen funds to a hidden externally owned account (EOA), says Alex Delamotte, the senior threat researcher with SentinelOne who wrote the analysis. ... The decentralized finance (DeFi) ecosystem relies on smart contracts — as well as other technologies such as blockchains, oracles, and key management — to execute transactions, manage data on a blockchain, and allow for agreements between different parties and intermediaries. Yet their linchpin status also makes smart contracts a focus of attacks and a key component of fraud. "A single vulnerability in a smart contract can result in the irreversible loss of funds or assets," Shashank says. "In the DeFi space, even minor mistakes can have catastrophic financial consequences. However, the danger doesn’t stop at monetary losses — reputational damage can be equally, if not more, damaging." ... Companies should take stock of all smart contracts by maintaining a detailed and up-to-date record of all deployed smart contracts, verifying every contract, and conducting periodic audits. Real-time monitoring of smart contracts and transactions can detect anomalies and provide fast response to any potential attack, says CredShields' Shashank.


Is AI the end of IT as we know it?

CIOs have always been challenged by the time, skills, and complexities involved in running IT operations. Cloud computing, low-code development platforms, and many DevOps practices helped IT teams move “up stack,” away from the ones and zeros, to higher-level tasks. Now the question is whether AI will free CIOs and IT to focus more on where AI can deliver business value, instead of developing and supporting the underlying technologies. ... Joe Puglisi, growth strategist and fractional CIO at 10xnewco, offered this pragmatic advice: “I think back to the days when you wrote an assembly and it took a lot of time. We introduced compilers, higher-level languages, and now we have AI that can write code. This is a natural progression of capabilities and not the end of programming.” The paradigm shift suggests CIOs will have to revisit their software development lifecycles for significant shifts in skills, practices, and tools. “AI won’t replace agile or DevOps — it’ll supercharge them with standups becoming data-driven, CI/CD pipelines self-optimizing, and QA leaning on AI for test creation and coverage,” says Dominik Angerer, CEO of Storyblok. “Developers shift from coding to curating, business users will describe ideas in natural language, and AI will build functional prototypes instantly. This democratization of development brings more voices into the software process while pushing IT to focus on oversight, scalability, and compliance.”


From Indicators to Insights: Automating Risk Amplification to Strengthen Security Posture

Security analysts don’t want more alerts. They want more relevant ones. Traditional SIEMs generate events using their own internal language that involve things like MITRE tags, rule names and severity scores. But what frontline responders really want to know is which users, systems, or cloud resources are most at risk right now. That’s why contextual risk modeling matters. Instead of alerting on abstract events, modern detection should aggregate risk around assets including users, endpoints, APIs, or services. This shifts the SOC conversation from “What alert fired?” to “Which assets should I care about today?” ... The burden of alert fatigue isn’t just operational but also emotional. Analysts spend hours chasing shadows, pivoting across tools, chasing one-off indicators that lead nowhere. When everything is an anomaly, nothing is actionable. Risk amplification offers a way to reduce the unseen yet heavy weight on security analysts and the emotional toll it can take by aligning high-risk signals to high-value assets and surfacing insights only when multiple forms of evidence converge. Rather than relying on a single failed login or endpoint alert, analysts can correlate chains of activity whether they be login anomalies, suspicious API queries, lateral movement, or outbound data flows – all of which together paint a much stronger picture of risk.


The Immune System of Software: Can Biology Illuminate Testing?

In software engineering, quality assurance is often framed as identifying bugs, validating outputs, and confirming expected behaviour. But similar to immunology, software testing is much more than verification. It is the process of defining the boundaries of the system, training it to resist failure, and learning from its past weaknesses. Like the immune system, software testing should be multi-layered, adaptive, and capable of evolving over time. ... Just as innate immunity is present from biological birth, unit tests should be present from the birth of our code. Just as innate immunity doesn't need a full diagnostic history to act, unit tests don’t require a full system context. They work in isolation, making them highly efficient. But they also have limits: they can't catch integration issues or logic bugs that emerge from component interactions. That role belongs to more evolved layers. ... Negative testing isn’t about proving what a system can do — it’s about ensuring the system doesn’t do what it must never do. It verifies how the software behaves when exposed to invalid input, unauthorized access, or unexpected data structures. It asks: Does the system fail gracefully? Does it reject the bad while still functioning with the good? Just as an autoimmune disease results from a misrecognition of the self, software bugs often arise when we misrecognise what our code should do and what it should not do.


CSO hiring on the rise: How to land a top security exec role

“Boards want leaders who can manage risk and reputation, which has made soft skills — such as media handling, crisis communication, and board or financial fluency — nearly as critical as technical depth,” Breckenridge explains. ... “Organizations are seeking cybersecurity leaders who combine technical depth, AI fluency, and strong interpersonal skills,” Fuller says. “AI literacy is now a baseline expectation, as CISOs must understand how to defend against AI-driven threats and manage governance frameworks.” ... Offers of top pay and authority to CSO candidates obviously come with high expectations. Organizations are looking for CSOs with a strong blend of technical expertise, business acumen, and interpersonal strength, Fuller says. Key skills include cloud security, identity and access management (IAM), AI governance, and incident response planning. Beyond technical skills, “power skills” such as communication, creativity, and problem-solving are increasingly valued, Fuller explains. “The ability to translate complex risks into business language and influence board-level decisions is a major differentiator. Traits such as resilience, adaptability, and ethical leadership are essential — not only for managing crises but also for building trust and fostering a culture of security across the enterprise,” he says.


From legacy to SaaS: Why complexity is the enemy of enterprise security

By modernizing, i.e., moving applications to a more SaaS-like consumption model, the network perimeter and associated on-prem complexity tends to dissipate, which is actually a good thing, as it makes ZTNA easier to implement. As the main entry point into an organization’s IT system becomes the web application URL (and browser), this reduces attackers’ opportunities and forces them to focus on the identity layer, subverting authentication, phishing, etc. Of course, a higher degree of trust has to be placed (and tolerated) in SaaS providers, but at least we now have clear guidance on what to look for when transitioning to SaaS and cloud: identity protection, MFA, and phishing-resistant authentication mechanisms become critical—and these are often enforced by default or at least much easier to implement compared to traditional systems. ... The unwillingness to simplify technology stack by moving to SaaS is then combined with a reluctant and forced move to the cloud for some applications, usually dictated by business priorities or even ransomware attacks (as in the BL case above). This is a toxic mix which increases complexity and reduces the ability for a resource-constrained organization to keep security risks at bay.


Why Metadata Is the New Interface Between IT and AI

A looming risk in enterprise AI today is using the wrong data or proprietary data in AI data pipelines. This may include feeding internal drafts to a public chatbot, training models on outdated or duplicate data, or using sensitive files containing employee, customer, financial or IP data. The implications range from wasted resources to data breaches and reputational damage. A comprehensive metadata management strategy for unstructured data can mitigate these risks by acting as a gatekeeper for AI workflows. For example, if a company wants to train a model to answer customer questions in a chatbot, metadata can be used to exclude internal files, non-final versions, or documents marked as confidential. Only the vetted, tagged, and appropriate content is passed through for embedding and inference. This is a more intelligent, nuanced approach than simply dumping all available files into an AI pipeline. With rich metadata in place, organizations can filter, sort, and segment data based on business requirements, project scope, or risk level. Metadata augments vector labeling for AI inferencing. A metadata management system helps users discover which files to feed the AI tool, such as health benefits documents in an HR chatbot while vector labeling gives deeper information as to what’s in each document.


Ask a Data Ethicist: What Should You Know About De-Identifying Data?

Simply put, data de-identification is removing or obscuring details from a dataset in order to preserve privacy. We can think about de-identification as existing on a continuum... Pseudonymization is the application of different techniques to obscure the information, but allows it to be accessed when another piece of information (key) is applied. In the above example, the identity number might unlock the full details – Joe Blogs of 123 Meadow Drive, Moab UT. Pseudonymization retains the utility of the data while affording a certain level of privacy. It should be noted that while the terms anonymize or anonymization are widely used – including in regulations – some feel it is not really possible to fully anonymize data, as there is always a non-zero chance of reidentification. Yet, taking reasonable steps on the de-identification continuum is an important part of compliance with requirements that call for the protection of personal data. There are many different articles and resources that discuss a wide variety of types of de-identification techniques and the merits of various approaches ranging from simple masking techniques to more sophisticated types of encryption. The objective is to strike a balance between the complexity of the the technique to ensure sufficient protection, while not being burdensome to implement and maintain.


5 ways business leaders can transform workplace culture - and it starts by listening

Antony Hausdoerfer, group CIO at auto breakdown specialist The AA, said effective leaders recognize that other people will challenge established ways of working. Hearing these opinions comes with an open management approach. "You need to ensure that you're humble in listening, but then able to make decisions, commit, and act," he said. "Effective listening is about managing with humility with commitment, and that's something we've been very focused on recently." Hausdoerfer told ZDNET how that process works in his IT organization. "I don't know the answer to everything," he said. "In fact, I don't know the answer to many things, but my team does, and by listening to them, we'll probably get the best outcome. Then we commit to act." ... Bev White, CEO at technology and talent solutions provider Nash Squared, said open ears are a key attribute for successful executives. "There are times to speak and times to listen -- good leaders recognize which is which," she said. "The more you listen, the more you will understand how people are really thinking and feeling -- and with so many great people in any business, you're also sure to pick up new information, deepen your understanding of certain issues, and gain key insights you need."


Beyond Efficiency: AI's role in reshaping work and reimagining impact

The workplace of the future is not about humans versus machines; it's about humans working alongside machines. AI's real value lies in augmentation: enabling people to do more, do better, and do what truly matters. Take recruitment, for example. Traditionally time-intensive and often vulnerable to unconscious bias, hiring is being reimagined through AI. Today, organisations can deploy AI to analyse vast talent pools, match skills to roles with precision, and screen candidates based on objective data. This not only reduces time-to-hire but also supports inclusive hiring practices by mitigating biases in decision-making. In fact, across the employee lifecycle, it personalises experiences at scale. From career development tools that recommend roles and learning paths aligned with individual aspirations, to chatbots that provide real-time HR support, AI makes the employee journey more intuitive, proactive, and empowering. ... AI is not without its challenges. As with any transformative technology, its success hinges on responsible deployment. This includes robust governance, transparency, and a commitment to fairness and inclusion. Diversity must be built into the AI lifecycle, from the data it's trained on to the algorithms that guide its decisions.