Daily Tech Digest by Kannan Subbiah: audit

Showing posts with label audit. Show all posts

Daily Tech Digest - August 17, 2026

Quote for the day:

"Listen with curiosity, speak with honesty act with integrity." -- Roy Bennett

🎧 Listen to the audio debrief on YouTube

Duration: 20 mins • Perfect for listening on the go.

How to level up from IT management to IT leadership

Transitioning from a mid-level technical management position to a senior executive role requires a deliberate shift in focus from mastering technology to mastering human connections and business operations. Aspiring leaders must build upon their foundational knowledge by developing essential communication habits, such as empathy, active listening, and the ability to build trust across different departments. Successfully navigating this career path involves taking on significant projects, learning from the inevitable missteps, and seeking out experienced mentors who can provide honest feedback. It is crucial to understand the broader goals of the organization and how technology can practically support those objectives. This means stepping away from the desk to learn about budgeting, risk management, and the daily challenges faced by other teams. True leadership is not defined by a specific title, but by the capacity to align people around a shared vision and empower them to succeed. Rather than simply executing technical tasks, effective leaders focus on mentoring their teams, translating complex concepts into plain language for non-technical coworkers, and making thoughtful decisions that deliver measurable value. Ultimately, ascending to the executive level is about solving company-wide problems with calm confidence and a steady collaborative mindset.

Why IoT systems fail at scale – and why Edge vs Cloud is the wrong debate

Internet of Things systems often struggle to scale, but the root cause is rarely the technology itself. Instead, failures usually stem from fragmented design. When teams develop hardware, software, connectivity, and security in isolation, the gaps between these components become major hurdles once the system moves into production. The ongoing debate pitting edge computing against the cloud misses the point. In practice, successful systems rely on both. The real challenge lies in deciding how they work together—specifically, figuring out which data should be processed locally for quick, time-sensitive tasks and which should be sent to the cloud for long-term analysis. This need for unified design is becoming even more obvious as artificial intelligence enters the picture. AI requires clear, reliable data pipelines. If a system's architecture is disjointed, having massive amounts of data won't help much. To build systems that last, developers need to shift from component-level thinking to holistic system design. This means planning data flow, security protocols, and long-term maintenance strategies from the very beginning. Treating features like security or software updates as add-ons only creates expensive problems later. By building a cohesive architecture from day one, organizations can create reliable systems that easily adapt and grow over time.

The new audit equation puts AI to work and judgement at the centre

In a recent interview, Atul Deshmukh of the accounting firm KNAV discusses how artificial intelligence is transforming the auditing profession from the ground up. Central to this shift is the transition from traditional statistical sampling to the comprehensive analysis of entire data sets. By deploying AI platforms, firms can automate repetitive and time-consuming tasks like document extraction and transaction matching. These digital workers drastically compress the time required for routine procedures, turning tasks that once took a full day into minutes. This efficiency is fundamentally altering the traditional accounting firm structure. The classic pyramid model, which relied heavily on junior staff for groundwork, is evolving into a diamond shape that demands analytical thinking and diverse backgrounds, including engineering. Furthermore, the massive time savings challenge the industry's conventional billable-hour model, paving the way for pricing based on value, complexity, and outcomes. Despite AI taking on larger segments of the workflow and even moving toward autonomous processes, human judgment remains the irreplaceable core of auditing. Auditors are not being replaced; their roles are shifting from manual verification to higher-level review and critical decision-making. Ultimately, AI handles the heavy lifting, allowing human professionals to focus their time on complex analysis and valuable insights.

What the CISO role will look like in 2029

By 2029, the role of the Chief Information Security Officer will shift away from being a purely technical position focused on building network defenses. Instead, security leaders will take on broader responsibilities as business strategists and risk managers. As technology cycles shorten and artificial intelligence accelerates the pace of both innovation and cyber threats, the old approach of simply saying no to all new ideas will no longer work. Tomorrow’s security executives will be expected to help their organizations take smart, calculated risks. Rather than managing security tools in isolation, future leaders will act as organizational orchestrators. They will connect engineering, legal, product, and executive teams to build systems that can identify and reduce risks almost instantly. Because threats are moving faster, organizations will rely on resilient engineering and automated decision-making processes to maintain safety. Some experts predict that the position will even expand to cover overall enterprise risk, potentially changing titles to emphasize trust and broader risk management. Despite these changes, the fundamental mission of the job remains steady. Security leaders will still need strong technical foundations, sound judgment, and clear communication skills to protect the entire business and help executives make informed choices in a rapidly changing world.

The Infrastructure Bottleneck That Keeps AI From Scaling Up

While many organizations focus entirely on choosing the right artificial intelligence models, the real challenge in making these systems work at a large scale lies in the underlying physical and technical foundational structures. According to Dilip Kumar of NTT DATA, practically all organizations find that their current networks, data storage, and security setups are slowing down their progress. Proving that an AI tool works in a small initial test is relatively simple, but running it reliably across an entire business is much harder. A common mistake is buying thousands of expensive software licenses without having the internal systems to actually use them. It is similar to buying a high-performance sports car but having no paved roads to drive it on. For AI to be truly useful, companies must ensure their networks can handle the data traffic and that their information is clean and organized. Instead of trying to transform an entire business at once, a smarter approach is to focus on a single, specific problem. By ensuring the foundation—the core networks, data organization, user identity, the appropriately sized model, and the daily operating procedures—is solid, businesses can prove the value of their investment quickly and then expand those efforts with complete confidence.

The Rise of Runtime Governance

In the article "The Rise of Runtime Governance," Christian Siegers argues that artificial intelligence forces a fundamental shift in how modern organizations manage system behavior. Historically, enterprise governance focused heavily on the implementation phase. Dedicated teams reviewed system architectures, assessed security measures, and validated strict compliance standards well before deployment. This approach was highly effective for traditional systems because their behavior was largely dictated by static code and predefined business rules. However, AI introduces a complex new dynamic where critical decisions actually occur during execution. Even if an AI system successfully passes all pre-deployment governance checks, its behavior can still drift due to changing context, model interactions, and new information retrieval. Consequently, companies may strictly follow governance processes without actually retaining control over the final operational outcomes. To bridge this gap, Siegers suggests that governance must evolve from a series of static checkpoints into a continuous architectural capability. This concept, known as runtime governance, requires embedding continuous system observability, active policy enforcement, and human oversight directly into the daily operational environment. By doing so, organizations can monitor what their systems are doing in real time, ensure all behavior remains within acceptable boundaries, and actively intervene when necessary. This ultimately maintains true control over AI-enabled operations long after the initial deployment.

Agentic Fitness Functions: Extending Evolutionary Architecture Beyond Deterministic Rules

Evolutionary software architecture relies on fitness functions—automated checks like dependency rules, performance budgets, and security scans—to ensure systems can change safely over time without degrading their core characteristics. While these deterministic rules are excellent for enforcing strict, measurable metrics, they often fall short when evaluating complex, judgment-heavy architectural concerns. For example, a basic schema check can confirm that an application programming interface still functions, but it cannot determine if a new field accidentally leaks user interface details into a core domain model. This is where agentic fitness functions come into play to fill the gap. By using artificial intelligence agents calibrated with past architectural decisions, ownership data, and clear rubrics, these functions can evaluate nuanced changes that defy simple yes-or-no rules. They are not meant to replace human architects or traditional automated tests. Instead, they act as an advisory layer that provides structured feedback, including confidence scores and clear reasoning, for changes that require context and human-like judgment. This approach helps teams maintain healthy system boundaries, catch semantic drift early, and ensure that architectural intent is preserved. Ultimately, agentic fitness functions make complex architectural decisions more transparent and auditable, allowing teams to confidently manage rapid software delivery and continuous system evolution.

From Agile to the Product Operating Model

Based on a recent survey of 48 practitioners, the transition from traditional development methods to a product operating model often changes company vocabulary and structure more than it changes how decisions are actually made. Among the respondents whose organizations are making this shift, most report that their teams still operate by building requested features rather than acting as fully empowered groups that decide how to solve problems. However, the survey does highlight some positive trends. Many participants notice improvements in the speed of delivery, the value provided to customers, and overall collaboration with stakeholders. On the other hand, business results remain largely inconclusive, likely because financial outcomes take longer to measure. One notable concern is the human element, as team morale and developer satisfaction appear to decline during these transitions. Additionally, the findings show that artificial intelligence adoption and structural operating changes are happening as separate efforts. While artificial intelligence is starting to influence how product decisions are made across many companies, this shift is occurring independently of formal organizational redesigns. Overall, the data suggests that while operational efficiency might improve, true changes in decision making authority and employee well being remain significant challenges for organizations attempting this transition today.

US cloud act, sovereignty, and why you might need to care

The article by Kate Carruthers discusses the crucial difference between data residency and true data sovereignty, emphasizing that physical location alone does not insulate data from foreign legal reach. Prompted by Airbus’s decision to move critical applications to a European provider, the piece highlights that the US CLOUD Act allows US authorities to compel American cloud providers to hand over data, regardless of whether that data is stored in Sydney, Frankfurt, or Dublin. This makes cloud hosting a matter of national security and governance, not just a technical or architectural choice. The author notes that Australia often mistakenly equates local data residency with sovereignty, creating a blind spot that leaves critical infrastructure vulnerable to geopolitical disputes or commercial shifts. Organizations are advised to map their vital dependencies and classify workloads based on the potential harm of disruption rather than blindly adopting a "cloud-first" strategy. Furthermore, companies should design systems for degraded operation, practice isolation techniques, and preserve clear exit options to ensure resilience. Ultimately, Carruthers argues that cloud computing has evolved into institutional and geopolitical infrastructure, requiring boards to make deliberate, strategic choices about where sensitive workloads sit and how much control they truly retain.

The cyber resilience divide

In today's digital landscape, security incidents are a routine reality, and companies can no longer rely solely on preventing attacks. A recent Fujitsu report explores the growing gap between organizations that successfully build strong defenses and those that remain vulnerable, particularly as artificial intelligence reshapes both security threats and defense strategies. While artificial intelligence helps criminals find weaknesses and automate attacks, it also provides companies with powerful tools to detect and respond to these threats early. The research identifies a clear division between leading organizations and those lagging behind. Leaders understand that security breaches are inevitable. Rather than focusing only on prevention, they prepare to maintain operations and recover quickly. They treat security as a shared priority that begins at the board level, balancing new technology adoption with careful oversight. By running practical simulations and using smart tools for defense, these leaders reduce the impact of incidents while building trust and supporting steady growth. In contrast, lagging organizations often rush to adopt new technologies without fully understanding the risks, leaving gaps in their defenses. To secure their futures, companies must accept that breaches will happen, embed security awareness into their daily routines, and focus on protecting their most important systems through practical testing.

Daily Tech Digest - May 23, 2026

Quote for the day:

“Great tech leadership isn’t about mastering every technology — it’s about creating the clarity and confidence for teams to build what doesn’t exist yet.” -- Anonymous

🎧 Listen to this digest on YouTube Music

▶ Play Audio Digest

Duration: 23 mins • Perfect for listening on the go.

Downtime has become a $600 billion business problem

According to Splunk's "The Hidden Costs of Downtime" report, unplanned outages and service degradations have escalated into a $600 billion problem for the Global 2000, representing a fifty percent surge over the last two years. Each affected organization experiences an average of sixty annual incidents, costing an average of $300 million per company. These mounting expenses include a near doubling of lost revenue to $95 million, alongside substantial climbs in regulatory fines to $51 million, driven by strict GDPR and DORA compliance enforcement, and ransomware payouts reaching $40 million. Beyond immediate financial blows, outages inflict severe long-term impacts, including delayed product launches, eroded brand trust that takes months to recover, and an average 3.4% stock value decline. The report highlights that third party dependencies, such as SaaS platforms and APIs, have become a primary catalyst for downtime, skyrocketing from 24% in 2024 to 63% in 2026, which severely hampers end to end infrastructure visibility. In response, enterprises are prioritizing visibility solutions and investing a median of $24.5 million annually into generative and agentic AI tools for rapid incident triage and root cause analysis. Geographically, EMEA faces the highest overall costs, while sector wise, information services and technology suffer the most severe impact at $402 million per company.

Making Vulnerable Drivers Exploitable Without Hardware - The BYOVD Perspective

The Hacker News article analyzes a method for bypassing hardware restrictions to interact with Windows kernel-mode drivers from user mode, specifically examining how this impacts driver-focused vulnerability research and Bring Your Own Vulnerable Driver (BYOVD) post-exploitation techniques. Vulnerable drivers are frequently weaponized by attackers to compromise system defenses, such as Endpoint Detection and Response (EDR) agents. However, many drivers developed for dedicated hardware are "hardware-gated," meaning they only instantiate their device objects or execute initialization routines (like AddDevice or IRP_MJ_PNP callbacks) if the corresponding hardware chip is detected. To assess exploitability in the absence of physical devices, researchers utilize userland-level deployment techniques that do not rely on standard kernel-mode debuggers or hardware virtualization. This includes using service creation commands like sc.exe to unconditionally load non-Plug and Play (PnP) drivers and evaluate whether named device objects are generated inside the \Devices directory. By mapping initialization logic and monitoring how the underlying PnP manager interacts with the driver extension, researchers can determine whether vulnerable paths, such as arbitrary memory read/write functions or Memory-Mapped I/O (MMIO) instructions, can be successfully reached and exploited entirely from userland with administrative privileges.

Leadership by Vibe Instead of Evidence

In her Medium article, Jodie Shaw examines the modern corporate tendency where executives treat personal confidence and gut instinct as strategic evidence, a phenomenon she terms "leadership by vibe." Shaw argues that while intuition is often culturally glorified, relying primarily on unchecked executive emotions or singular observations creates organizational volatility, erodes worker trust, and prompts teams to manage their leaders' feelings rather than actual performance. Citing a variety of research, she highlights how power distorts perception, causing executive confidence to outpace factual accuracy and forcing discouraged employees to view corporate strategy as merely temporary. This persistent reliance on unverified assumptions yields devastating real-world financial and operational outcomes, such as Peloton’s catastrophic pandemic forecasting errors that triggered massive quarterly losses, and the BBC’s holiday pay scandal that cost over £300 million due to unchallenged institutional memories. To counteract this operational drift, Shaw points to data-driven organizations like Toyota, Shopify, and Netflix. These forward-thinking companies intentionally implement robust structural constraints, such as firsthand observations, automated kill metrics, and team pre-mortems, to reframe intuition as a mere hypothesis rather than an infallible plan. Ultimately, true leadership demands the humility to confront uncomfortable data and prioritize evidence over emotional reactivity.

The Hidden Cost of Bad Data: Financial Institutions Lose Millions Without Knowing It

In this article, Gayathri Balakumar, a lead data engineer at Capital One, argues that financial institutions bleed substantial capital not from market conditions, but because they have normalized the dysfunction of poor data quality. This silent crisis often goes unnoticed because its financial toll does not appear as a distinct line item on profit and loss statements. Instead, it severely compromises credit decisions, delays operational flows, and results in missed market opportunities. McKinsey and Company estimates that bad data inflates banking operational costs by 15% to 25%. Furthermore, banks cannot successfully deploy advanced technologies like artificial intelligence or digital transformations if their underlying foundation remains structurally compromised, fragmented, or outdated. Rather than investing heavily in downstream damage control, such as manual reconciliations, duplicate databases, and post-processing validation teams, bank leaders must treat data as a critical strategic asset. Balakumar advocates for a proactive leadership mandate focusing on real-time integration, unified architectures, strict data ownership, and the deployment of autonomous agentic AI frameworks to clean and standardize information at the point of entry. Ultimately, financial institutions that directly confront these systemic inefficiencies will eliminate massive hidden costs, accurately forecast market risks, and secure a lasting competitive edge over rivals who continue to patch over flaws.

Everyone Suddenly Wants Claude's Audit Logs

The article reports that 27 enterprise security vendors have announced integrations with Anthropic's Claude Compliance API to manage the platform's activity data inside corporate security environments. Initially launched in August 2025, the structured API feed eliminates manual log exports by programmatically feeding real-time user behavior, login activity, and administrative shifts into preexisting enterprise monitoring setups. For Claude Enterprise users, the data includes specific conversational content and uploaded files, which is crucial given data showing that 4% of prompts leak private information and 20% of uploaded files contain confidential information. Major vendors like Cloudflare, CrowdStrike, and Microsoft are integrating this API into their respective stacks to handle threat detection, automated incident response, and unified AI governance across multiple assistants. This massive vendor alignment stems from a dramatic rise in enterprise adoption of Claude, which escalated from 56.2% to 94.9% between April 2025 and April 2026. However, industry experts caution that executing the Compliance API represents only "half a story" for highly regulated industries. Because the tool manages control plane data rather than localized network-layer inputs or agent-level operational workflows, organizations must implement additional telemetry to ensure complete corporate audit coverage.

Architects Are Not Here to Keep the Lights On

In this article, Paul Preiss disputes the common executive misconception that IT architects exist merely to manage existing technology estates, handle portfolio rationalization, or ensure basic operational continuity. Instead, utilizing the Business Technology Architecture Body of Knowledge (BTABoK) framework, Preiss asserts that the entire architectural profession is fundamentally oriented around driving innovation, managing transformation, and delivering new business value through proactive strategy. This change-focused approach applies across all five recognized specializations: business architects bridge strategy and technical delivery; software architects make structural decisions within active deployment; information architects transform data into a genuine lever for competitive disruption; infrastructure architects engineer the broad compute landscapes of the future; and solution architects orchestrate delivery across programs, products, and projects. Furthermore, the text advocates for a chief architect model where senior leaders maintain active, hands-on delivery responsibilities, which is analogous to a chief of medicine continuing to treat patients, rather than drifting into detached, purely administrative management positions that lose technical competency. Ultimately, the architectural lifecycle continuously loops through measurement to build the evidence base for subsequent transformations. Rather than preserving past investments, architects must act as genuine change agents within complex corporate ecosystems to maximize organizational velocity, reduce deployment risks, and secure long-term digital advantages.

The sovereign cloud illusion

In this InfoWorld opinion piece, technology expert David Linthicum argues that the concept of a sovereign cloud is largely a marketing illusion rather than a realistic, off-the-shelf procurement option. True digital sovereignty demands absolute independence across a full hardware and software stack, which encompasses local data residency, platform ownership, codebase control, chip manufacturing, regular software patching, and clear legal jurisdiction. In practical terms, only the United States and China currently possess the immense scale, global engineering depth, and operational maturity required to sustain these entirely independent infrastructures. Consequently, regional European initiatives such as Gaia-X, Andromeda, and Numergy have historically struggled to achieve lasting competitive gravity against deeply consolidated American hyperscalers. Even when localized regions are deployed by dominant global vendors, they inherently retain dependencies on external parent companies and remote control planes that effectively phone home. Rather than fruitlessly chasing an unattainable ideal or mistakenly adopting unportable multicloud architectures, Linthicum advises enterprise leaders to view cloud sovereignty as a broad spectrum of risk reduction choices. Organizations must accurately audit existing dependencies, isolate sensitive enterprise workloads, minimize reliance on proprietary platform features, and implement robust, fully funded exit strategies to insulate themselves from future geopolitical conflicts.

Valid certificates, stolen accounts: how attackers broke npm's last trust signal

The VentureBeat article details how a major supply chain attack compromised 633 malicious npm package versions, enabling them to bypass Sigstore provenance verification by leveraging stolen OpenID Connect tokens from legitimate maintainer accounts. Because Sigstore only validates that a package originates from a continuous integration environment without confirming explicit publisher authorization, this incident highlights a severe vulnerability in automated trust signals. This breach is part of a broader trend exposing seven critical developer tool attack surfaces, including VS Code extension credential theft, Model Context Protocol server automated execution, continuous integration agent prompt injection, agent framework code execution, IDE credential storage vulnerabilities, and shadow AI exposure. Security research shows that popular AI coding command line interfaces automatically execute untrusted local configurations, and prompt injections can trick AI agents into leaking sensitive API keys. Crucially, adversaries are actively exploiting these gaps to hunt for personal access tokens, cloud credentials, and corporate source code. To counter these invisible blind spots that traditional endpoint detection and data loss prevention systems cannot monitor, the article provides a specialized audit grid. It strongly recommends that organizations implement dual party publication approvals for packages, enforce strict minimum age policies for extension updates, and establish browser layer AI governance to robustly protect infrastructure intelligence from sophisticated identity theft.

How concerned should CIOs be with geopolitics?

According to the CIO article, growing global tensions and sophisticated cyber threats have elevated digital and technological sovereignty to a top strategic priority for enterprise boards and IT leaders. This shift has prompted a major emphasis on where technology is built and operated to reduce critical dependencies on third-party countries. According to Deloitte's Manel Barahona, 77% of organizations now view a provider's country of origin as a decisive factor, shifting focus beyond mere cost or performance toward business continuity and risk mitigation. This trend is driving massive financial commitments; Forrester projects that European investments in AI, cloud, and data sovereignty technologies will rise by 6.3% to a record €1.5 trillion. To navigate these geopolitical uncertainties, progressive CIOs like David Marimón of Coca-Cola European Partners and Álvaro Ontañón of Merlin Properties advocate for pragmatic strategies that balance day-to-day operational efficiency with long-term resilience. Consequently, organizations are actively diversifying suppliers, designing hybrid architectures to maintain strategic optionality, and evaluating local and regional capabilities. This landscape has transformed the CIO role into a highly cross-functional, decisive boardroom position tasked with managing technological dependence as a primary strategic risk while aligning infrastructure directly with legal frameworks, corporate values, and overall business competitiveness.

The Data Analytics Fallacies Your Team Is Treating as Best Practices

The Dataversity article explores insidious data analytics fallacies that modern teams frequently mistake for industry best practices, creating polished dashboards built on flawed assumptions. The author highlights five central traps that compromise strategic decisions. First, correlation often drives organizational decisions under the guise of causation, prompting misguided budget shifts or product modifications without an understanding of the underlying operational mechanisms. Second, survivorship bias frequently masquerades as insight, causing teams to analyze a highly filtered reality of successful outcomes while ignoring vital context from failed experiments or churned users. Third, over-engineered metrics provide a false sense of comfort, burying minor, unverified statistical assumptions inside complex formulas that operate entirely on unearned trust. Fourth, incomplete sampling creates a misleading illusion of completeness, confining teams to narrow dataset slices while leaving broader structural realities unaddressed. Finally, confirmation bias subtly embeds itself within analytical processes as queries are iteratively refined to align with preexisting management expectations, often resulting in the systematic deletion of inconvenient outliers. Ultimately, the piece warns that the most dangerous analytical mistakes appear highly structured and persuasive, urging organizations to critically evaluate the core logic behind their metrics rather than blindly accepting polished visual reports.

Daily Tech Digest - February 24, 2026

Quote for the day:

"Transparent reviews create fairness. Subjective reviews create frustration." -- Gordon Tredgold

AI agents and bad productivity metrics

The great promise of was that it would finally clear our backlogs. Coding agents would churn out boilerplate at superhuman speeds, and teams would finally ship exactly what the business wants. The reality, as we settle into 2026, is far more uncomfortable. Artificial intelligence is not going to save developer productivity because writing code was never the bottleneck in software engineering. ... For decades, one of the most common debugging techniques was entirely social. A production alert goes off. You look at the version control history, find the person who wrote the code, ask them what they were trying to accomplish, and reconstruct the architectural intent. But what happens to that workflow when no one actually wrote the code? What happens when a human merely skimmed a 3,000-line agent-generated pull request, hit merge, and moved on to the next ticket? When an incident happens, where is the deep knowledge that used to live inside the author? ... The metrics that matter are still the boring ones because they measure actual business outcomes. The remain the best sanity check we have because they tie delivery speed directly to system stability. They measure deployment frequency, lead time for changes, change failure rate, and time to restore service. None of those metrics cares about the number of commits your agents produced today. They only care about whether your system can absorb change without breaking.

How vertical SaaS is redefining enterprise efficiency

For the past decade, horizontal SaaS has been the defining force in enterprise technology. Platforms like CRMs, ERP suites and collaboration tools promised universality, offering a single platform to manage every business function across all industries. The strategy made sense: a large total addressable market, reusable architecture and marketing scale. flips that model. It is narrow by design but deep in impact. A report by Strategy& found that B2B vertical software companies are now growing faster than their horizontal peers, thanks to higher retention rates, lower churn rates and better unit economics. When software mirrors how a business already works, people stop treating it like a tool they tolerate and start relying on it like infrastructure. ... In regulated industries, isn’t a feature; it’s the baseline for trust. I learned early that trying to retrofit audit trails or data retention policies after go-live only creates technical debt. Instead, design for compliance as a first-class product layer: immutable logs, permission hierarchies and exportable compliance reports built into the system. ... Vertical products don’t thrive in isolation. Integration with industry hardware, marketplaces and regulatory systems drives adoption. In one case, we partnered with a hardware vendor to automatically sync manifest data from their devices, cutting onboarding time in half and unlocking co-marketing opportunities.

API Security Standards: 10 Essentials to Get You Started

Most API security flaws are created during the design phase. You're too late if you're waiting until deployment to think about threats. Shift-left principles mean integrating security early, especially at the design phase, where flawed assumptions become future exploits. Start by mapping out each endpoint's purpose, what data it touches, and who should access it. Identify where trust is assumed (not earned), roles blur, and inputs aren't validated. ... Every API has a breaking point. If you don't define it, attackers will. prevent denial-of-service (DoS) attacks, and they're also your first defense against scraping, brute-forcing, enumeration, and even accidental misuse by poorly built integrations. APIs, by nature, invite automation. Without guardrails, that openness turns into a floodgate. And in some cases, unchecked abuse opens the door to far worse issues, like remote code execution, where improperly scoped input or lack of throttling leads directly to exploitation. ... APIs are built to accept input. Attackers find ways to exploit it. The core rule is this - if you didn't expect it, don't process it. If you didn't define it, don't send it. Define request and response schemas explicitly using tools like or JSON Schema, as recommended by leading API security standards. Then enforce them — at the gateway, app layer, or both. Don't just use validation as linting; treat it as a runtime contract. If the payload doesn't match the spec, reject it.

Why AI Urgency Is Forcing a Data Governance Reset

The cost of weak governance shows up in familiar ways: teams can’t find data, requirements arrive late in the process, and launches stall when compliance realities collide with product timelines. Without governance, argues, organizations “ultimately suffer from higher cost basis,” with downstream consequences that “impact the bottom line.” ... McQuillan sees a clear step-change in executive urgency since generative AI (GenAI) became mainstream. “There’s been a rapid adoption, particularly since the advent of GenAI and the type of generative and agentic technologies that a lot of C-suites are taking on,” he says. But he also describes a common leadership gap: many executives feel pressure to become “AI-enabled” without a clear definition of what that means or how to build it sustainably. “There’s very much a well-understood need across all companies to become AI-enabled in some way,” he says. “But the problem is a lot of folks don’t necessarily know how to define that.” In the absence of clarity, organizations often fall into scattershot experimentation. What concerns McQuillan the most is how the pace of the “race” shapes priorities. ... When asked whether the long-running mantra “data is the new oil” still holds in the era of large language models and agentic workflows, McQuillan is direct. “It holds true now more than ever,” he says. He acknowledges why attention drifts: “It’s natural for people to gravitate toward things that are shiny,” and “AI in and of itself is an absolutely magnificent space.”

Building a Least-Privilege AI Agent Gateway for Infrastructure Automation with MCP, OPA, and Ephemeral Runners

An agent misinterpreting an instruction can initiate destructive infrastructure changes, such as tearing down environments or modifying production resources. A compromised agent identity can be abused to exfiltrate secrets, create unauthorized workloads, or consume resources at scale. In practice, teams often discover these issues late, because traditional logs record what happened, but not why an agent decided to act in the first place. For organizations, this liability creates operational and governance challenges. Incidents become harder to investigate, change approvals are bypassed unintentionally, and security teams are left with incomplete audit trails. Over time, this problem erodes trust in automation itself, forcing teams to either roll back agent usage or accept increasing levels of unmanaged risk. ... A more sustainable approach is to introduce an explicit control layer between agents and the systems they operate on. In this article, we focus on an AI Agent Gateway, a dedicated boundary that validates intent, enforces policy as code, and isolates execution before any infrastructure or service API is invoked. Rather than treating agents as privileged actors, this model treats them as untrusted requesters whose actions must be authorized, constrained, observed, and contained. ... In the context of AI-driven automation, defense in depth means that no single component, neither the agent, nor the gateway, nor the execution environment, has enough authority on its own to cause damage.

Demystifying CERT‑In’s Elemental Cyber Defense Controls: A Guide for MSMEs

For India’s Micro, Small, and Medium Enterprises (MSMEs), cybersecurity is no longer a “big company problem.” With digital payments, SaaS adoption, cloud-first operations, and supply‑chain integrations becoming the norm, MSMEs are now prime targets for cyberattacks. To help these organizations build a strong foundational security posture, the Indian Computer Emergency Response Team (CERT-In) has released CIGU-2025-0003, outlining a baseline of Cyber Defense Controls, which prescribes 15 Elemental Cyber Security Controls—a pragmatic, baseline set of safeguards designed to uplift the nation’s cyber hygiene. ... These controls, mapped to 45 recommendations, enable essential digital hygiene, protect against ransomware, ensure regulatory compliance, and are required for annual audits. CERT‑In’s Elemental Controls are designed as minimum essential practices that every Indian organization—regardless of size—should implement. ... The CERT-In guidelines offer a simplified, actionable starting point for MSMEs to benchmark their security. These controls are intentionally prescriptive, unlike ISO or NIST, which are more framework‑oriented. ... Because threats constantly evolve and MSMEs face unique risks depending on their industry and data sensitivity, organizations should view this framework not as an endpoint, but as the first critical step toward building a comprehensive security program akin to ISO 27001 or NIST CSF 2.0.

AI-fuelled cyber attacks hit in minutes, warns CrowdStrike

CrowdStrike reports a sharp acceleration in cyber intrusions, with attackers moving from initial access to lateral movement in less than half an hour on average as widely available artificial intelligence tools become embedded in criminal workflows. Its latest Global Threat Report puts average eCrime "breakout time" at 29 minutes in 2025, a 65% improvement on the prior year. ... Alongside generative AI use in preparation and execution, the report describes attempts to exploit AI systems directly. Adversaries injected malicious prompts into GenAI tools at more than 90 organisations, using them to generate commands associated with credential theft and cryptocurrency theft. ... Incidents linked to North Korea rose more than 130%, while activity by the group CrowdStrike tracks as FAMOUS CHOLLIMA more than doubled. The report says DPRK-nexus actors used AI-generated personas to scale insider operations. It also cites a large cryptocurrency theft attributed to the actor it calls PRESSURE CHOLLIMA, valued at USD $1.46 billion and described as the largest single financial heist ever reported. The report also references AI-linked tooling used by other state and criminal groups. Russia-nexus FANCY BEAR deployed LLM-enabled malware, which it named LAMEHUG, for automated reconnaissance and document collection. The eCrime actor tracked as PUNK SPIDER used AI-generated scripts to speed up credential dumping and erase forensic evidence.

Shadow mode, drift alerts and audit logs: Inside the modern audit loop

When systems moved at the speed of people, it made sense to do compliance checks every so often. But AI doesn't wait for the next review meeting. The change to an inline audit loop means audits will no longer occur just once in a while; they happen all the time. Compliance and risk management should be "baked in" to the AI lifecycle from development to production, rather than just post-deployment. This means establishing live metrics and guardrails that monitor AI behavior as it occurs and raise red flags as soon as something seems off. ... Cultural shift is equally important: Compliance teams must act less like after-the-fact auditors and more like AI co-pilots. In practice, this might mean compliance and AI engineers working together to define policy guardrails and continuously monitor key indicators. With the right tools and mindset, real-time AI governance can “nudge” and intervene early, helping teams course-correct without slowing down innovation. In fact, when done well, continuous governance builds trust rather than friction, providing shared visibility into AI operations for both builders and regulators, instead of unpleasant surprises after deployment. ... Shadow mode is a way to check compliance in real time: It ensures that the model handles inputs correctly and meets policy standards before it is fully released. One AI security framework showed how this method worked: Teams first ran AI in shadow mode, then compared AI and human inputs to determine trust.

Making AI Compliance Practical: A Guide for Data Teams Navigating Risk, Regulation, and Reality

As AI tools become more embedded in enterprise workflows, data teams are encountering a growing reality: compliance isn’t only a legal concern but also a design constraint, a quality signal, and, often, a competitive differentiator. But navigating compliance can feel complex, especially for teams focused on building and shipping. What is the good news? It doesn’t have to be. When approached intentionally, compliance becomes a pathway to better decisions, not a barrier. ... Automation can help with regulations, but only if it's used correctly. I've looked at a tool before that used algorithms to find private information. It worked well with English, but when tested with material in more than one language, it missed a few personal identifiers. The group thought it was "smart enough." It wasn't. We kept the automation, but we added human review for rare cases, confidence levels to make checks happen, and alerts for input formats that aren't common. The automation stayed the same, but there were built-in checks and balances. ... The biggest compliance failures don’t come from bad people. They come from good teams moving fast, skipping hard questions, and assuming nothing will go wrong. But compliance isn’t a blocker. It’s a product quality signal. People will trust you more if they are aware that your team has carefully considered the details.

Tata Communications’ Andrew Winney on why SASE is now non-negotiable

is often discussed as a product decision, but in reality it is a journey. Many enterprises start with a few use cases, such as securing internet access or enabling remote access to private applications. But they do not always extend those principles across contractors, third-party users, software-as-a-service applications and hybrid environments. Practical Zero Trust requires enterprises to rethink access fundamentally. Every request must be evaluated based on who the user is, the context from which they are accessing, the device they are using and the resource they are requesting. Access must then be granted only to that specific resource. ... represents a structural convergence of networking and security rather than a simple technology swap. What are the most critical architectural and change-management considerations enterprises must address during this transition? SASE is not a one-time technology change. It represents the convergence of networking and security under unified orchestration and policy management. That transition takes time and must be managed carefully. We typically work with enterprises through phased transition plans. If an organisation’s immediate priority is securing internet access or private application access for remote users, we begin there and expand to additional use cases over time. Integration is critical. Enterprises have existing investments in cloud platforms, local area networks and security tools.

Daily Tech Digest - July 26, 2025

Quote for the day:

"Small daily improvements over time lead to stunning results." -- Robin Sharma

Data Engineering in the Age of AI: Skills To Master Now

Streaming requires a new mindset. You must reason about event time compared to processing time, manage watermarking and windowing and guarantee exactly-once semantics even when things change midstream. These design patterns must be built into your pipelines from the beginning. ... Agentic AI stretches the typical data engineer’s streaming data skill set because it is no longer about a single model running in isolation. Today, we see networks of perception agents, reasoning agents and execution agents working together, each handling tasks and passing insights to the next in real time. If you know only how to schedule batch ETL jobs or deploy an inference server, you’re missing a core skill: how to build high-throughput, low-latency pipelines that keep these agents reliable and responsive in production. ... A single slow or broken stream can cause cascading failures in multiagent systems. Use schema registries, enforce data contracts and apply exactly-once semantics to maintain trust in your streaming infrastructure. ... Communication presents another challenge. Data scientists often discuss “precision” as a metric that data engineers must translate into reality. Implement evaluation scores like factual consistency checks, entity precision comparisons and human-in-the-loop review pipelines.

Anthropic unveils ‘auditing agents’ to test for AI misalignment

Anthropic said the first environment it developed tests an agent’s ability to complete an alignment audit for an intentionally misaligned model. For this test, the researchers brought the investigator agent to the environment and equipped it with similar tools that human teams use. The researchers instructed the agent to identify the root cause of the misalignment and determine the specific behaviors that were misaligned. According to the researchers, the investigator agent successfully found the root cause of the issues 10-13% of the time. Still, this performance improved to 42% “when using a super-agent approach that aggregates findings across multiple investigations.” ... The next step in most alignment audits is typically a behavioral evaluation, which determines whether the model consistently misbehaves in various settings. To test its evaluation agent, the researchers designed 10 models “given system prompt instructions to exhibit a specific behavior, such as excessive deference to the user.” They ran the agent five times per model and saw that the agent correctly finds and flags at least one quirk of the model. However, the agent sometimes failed to identify specific quirks consistently. It had trouble evaluating subtle quirks, such as self-promotion and research-sandbagging, as well as quirks that are difficult to elicit, like the Hardcode Test Cases quirk.

The agentic experience: Is MCP the right tool for your AI future?

As enterprises race to operationalize AI, the challenge isn't only about building and deploying large language models (LLMs), it's also about integrating them seamlessly into existing API ecosystems while maintaining enterprise level security, governance, and compliance. Apigee is committed to lead you in this journey. Apigee streamlines the integration of gen AI agents into applications by bolstering their security, scalability, and governance. While the Model Context Protocol (MCP) has emerged as a de facto method of integrating discrete APIs as tools, the journey of turning your APIs into these agentic tools is broader than a single protocol. This post highlights the critical role of your existing API programs in this evolution and how ... Leveraging MCP services across a network requires specific security constraints. Perhaps you would like to add authentication to your MCP server itself. Once you’ve authenticated calls to the MCP server you may want to authorize access to certain tools depending on the consuming application. You may want to provide first class observability information to track which tools are being used and by whom. Finally, you may want to ensure that whatever downstream APIs your MCP server is supplying tools for also has minimum guarantees of security like already outlined above

AI Innovation: 4 Steps For Enterprises To Gain Competitive Advantage

A skill is a single ability, such as the ability to write a message or analyze a spreadsheet and trigger actions from that analysis. An agent independently handles complex, multi-step processes to produce a measurable outcome. We recently announced an expanded network of Joule Agents to help foster autonomous collaboration across systems and lines of business. This includes out-of-the-box agents for HR, finance, supply chain, and other functions that companies can deploy quickly to help automate critical workflows. AI front-runners, such as Ericsson, Team Liquid, and Cirque du Soleil, also create customized agents that can tackle specific opportunities for process improvement. Now you can build them with Joule Studio, which provides a low-code workspace to help design, orchestrate, and manage custom agents using pre-defined skills, models, and data connections. This can give you the power to extend and tailor your agent network to your exact needs and business context. ... Another way to become an AI front-runner is to tackle fragmented tools and solutions by putting in place an open, interoperable ecosystem. After all, what good is an innovative AI tool if it runs into blockers when it encounters your other first- and third-party solutions?

Hard lessons from a chaotic transformation

The most difficult part of this transformation wasn’t the technology but getting people to collaborate in new ways, which required a greater focus on stakeholder alignment and change management. So my colleague first established a strong governance structure. A steering committee with leaders from key functions like IT, operations, finance, and merchandising met biweekly to review progress and resolve conflicts. This wasn’t a token committee, but a body with authority. If there were any issues with data exchange between marketing and supply chain, they were addressed and resolved during the meetings. By bringing all stakeholders together, we were also able to identify discrepancies early on. For example, when we discovered a new feature in the inventory system could slow down employee workflows, the operations manager reported it, and we immediately adjusted the rollout plan. Previously, such issues might not have been identified until after the full rollout and subsequent finger-pointing between IT and business departments. The next step was to focus on communication and culture. From previous failed projects, we knew that sending a few emails wasn’t enough, so we tried a more personal approach. We identified influential employees in each department and recruited them as change champions.

Benchmarks for AI in Software Engineering

HumanEval and SWE-bench have taken hold in the ML community, and yet, as indicated above, neither is necessarily reflective of LLMs’ competence in everyday software engineering tasks. I conjecture one of the reasons is the differences in points of view of the two communities! The ML community prefers large-scale, automatically scored benchmarks, as long as there is a “hill climbing” signal to improve LLMs. The business imperative for LLM makers to compete on popular leaderboards can relegate the broader user experience to a secondary concern. On the other hand, the software engineering community needs benchmarks that capture specific product experiences closely. Because curation is expensive, the scale of these benchmarks is sufficient only to get a reasonable offline signal for the decision at hand (A/B testing is always carried out before a launch). Such benchmarks may also require a complex setup to run, and sometimes are not automated in scoring; but these shortcomings can be acceptable considering a smaller scale. For exactly these reasons, these are not useful to the ML community. Much is lost due to these different points of view. It is an interesting question as to how these communities could collaborate to bridge the gap between scale and meaningfulness and create evals that work well for both communities.

Scientists Use Cryptography To Unlock Secrets of Quantum Advantage

When a quantum computer successfully handles a task that would be practically impossible for current computers, this achievement is referred to as quantum advantage. However, this advantage does not apply to all types of problems, which has led scientists to explore the precise conditions under which it can actually be achieved. While earlier research has outlined several conditions that might allow for quantum advantage, it has remained unclear whether those conditions are truly essential. To help clarify this, researchers at Kyoto University launched a study aimed at identifying both the necessary and sufficient conditions for achieving quantum advantage. Their method draws on tools from both quantum computing and cryptography, creating a bridge between two fields that are often viewed separately. ... “We were able to identify the necessary and sufficient conditions for quantum advantage by proving an equivalence between the existence of quantum advantage and the security of certain quantum cryptographic primitives,” says corresponding author Yuki Shirakawa. The results imply that when quantum advantage does not exist, then the security of almost all cryptographic primitives — previously believed to be secure — is broken. Importantly, these primitives are not limited to quantum cryptography but also include widely-used conventional cryptographic primitives as well as post-quantum ones that are rapidly evolving.

It’s time to stop letting our carbon fear kill tech progress

With increasing social and regulatory pressure, reluctance by a company to reveal emissions is ill-received. For example, in Europe the Corporate Sustainability Reporting Directive (CSRD) currently requires large businesses to publish their emissions and other sustainability datapoints. Opaque sustainability reporting undermines environmental commitments and distorts the reference points necessary for net zero progress. How can organisations work toward a low-carbon future when its measurement tools are incomplete or unreliable? The issue is particularly acute regarding Scope 3 emissions. Scope 3 emissions often account for the largest share of a company’s carbon footprint and are those generated indirectly along the supply chain by a company’s vendors, including emissions from technology infrastructure like data centres. ... It sounds grim, but there is some cause for optimism. Most companies are in a better position than they were five years ago and acknowledge that their measurement capabilities have improved. We need to accelerate the momentum of this progress to ensure real action. Earth Overshoot Day is a reminder that climate reporting for the sake of accountability and compliance only covers the basics. The next step is to use emissions data as benchmarks for real-world progress.

Why Supply Chain Resilience Starts with a Common Data Language

Building resilience isn’t just about buying more tech, it’s about making data more trustworthy, shareable, and actionable. That’s where global data standards play a critical role. The most agile supply chains are built on a shared framework for identifying, capturing, and sharing data. When organizations use consistent product and location identifiers, such as GTINs (Global Trade Item Numbers) and GLNs (Global Location Numbers) respectively, they reduce ambiguity, improve traceability, and eliminate the need for manual data reconciliation. With a common data language in place, businesses can cut through the noise of siloed systems and make faster, more confident decisions. ... Companies further along in their digital transformation can also explore advanced data-sharing standards like EPCIS (Electronic Product Code Information Services) or RFID (radio frequency identification) tagging, particularly in high-volume or high-risk environments. These technologies offer even greater visibility at the item level, enhancing traceability and automation. And the benefits of this kind of visibility extend far beyond trade compliance. Companies that adopt global data standards are significantly more agile. In fact, 58% of companies with full standards adoption say they manage supply chain agility “very well” compared to just 14% among those with no plans to adopt standards, studies show.

Opinion: The AI bias problem hasn’t gone away you know

When we build autonomous systems and allow them to make decisions for us, we enter a strange world of ethical limbo. A self-driving car forced to make a similar decision to protect the driver or a pedestrian in a case of a potentially fatal crash will have much more time than a human to make its choice. But what factors influence that choice? ... It’s not just the AI systems shaping the narrative, raising some voices while quieting others. Organisations made up of ordinary flesh-and-blood people are doing it too. Irish cognitive scientist Abeba Birhane, a highly-regarded researcher of human behaviour, social systems and responsible and ethical artificial intelligence was asked to give a keynote recently for the AI for Good Global Summit. According to her own reports on Bluesky, a meeting was requested just hours before presenting her keynote: “I went through an intense negotiation with the organisers (for over an hour) where we went through my slides and had to remove anything that mentions ‘Palestine’ ‘Israel’ and replace ‘genocide’ with ‘war crimes’…and a slide that explains illegal data torrenting by Meta, I also had to remove. In the end, it was either remove everything that names names (Big Tech particularly) and remove logos, or cancel my talk.”

Daily Tech Digest - March 13, 2025

Quote for the day:

"Your present circumstances don’t determine where you can go; they merely determine where you start." -- Nido Qubein

Becoming an AI-First Organization: What CIOs Must Get Right

"The three pillars of an AI-first organization are data, infrastructure and people. Data must be treated as a strategic asset with robust quality, privacy and security standards," Simha said. Along with responsible AI, responsible data management is equally crucial. When implemented effectively, data privacy, regulatory compliance, bias and security do not pose issues to an AI-first organization. Yeo described the AI-first approach as both a journey and a destination. "Just using AI tools doesn't make you AI-first. Organizations must explore AI's full potential." He compared today's AI evolution to the early days of the internet. "Decades ago, businesses knew they had to go online but didn't know how. Now, if you're not online, you're obsolete. AI is following the same trajectory - it will soon be indispensable for business success." ... Simha stressed the importance of enterprise architecture in AI deployment. "AI success depends on how well data flows across an organization. Organizations must select the right architecture patterns - real-time data processing requires a Kappa architecture, while periodic reporting benefits from a Lambda approach. A well-designed data foundation is crucial," Simha said. As AI adoption grows, ethical concerns and regulatory compliance remain critical considerations.

From Box-Ticking to Risk-Tackling: Evolving Your GRC Beyond Audits

The problem, though, is that merely passing an audit does not necessarily mean a business is doing all it can to mitigate its risks. On their own, audits can fall short of driving full GRC maturity for several reasons ... Auditors are generally outsiders to the businesses they audit — which is good in the sense that it makes them objective evaluators. But it can also lead to situations where they have a limited understanding of what's really going on within a company's GRC practices and are beholden to the information provided by the company's team members on the other side of the assessment table. They may not ask the questions needed to gain adequate understanding to assess and find gaps, ultimately overlooking pitfalls that only insiders know about, and which would become obvious only following a higher degree of scrutiny than a standardized audit. ... But for companies that have made advanced GRC investments, such as automations that pull data from across a diverse set of disparate systems, deeper scrutiny will help validate the value that these investments have created. It may also uncover risk management weak points that the business is overlooking, allowing it to strengthen its GRC program even further. It's generally OK, by the way, if your business submits itself to a high degree of risk management scrutiny, only to fail the assessment because its controls are not as robust as it expected.

How to use ChatGPT to write code - and my favorite trick to debug what it generates

After repeated tests, it became clear that if you ask ChatGPT to deliver a complete application, the tool will fail. A corollary to this observation is that if you know nothing about coding and want ChatGPT to build something, it will fail. Where ChatGPT succeeds -- and does so very well -- is in helping someone who already knows how to code to build specific routines and get tasks done. Don't ask for an app that runs on the menu bar. But if you ask ChatGPT for a routine to put a menu on the menu bar, and paste that into your project, the tool will do quite well. Also, remember that, while ChatGPT appears to have a tremendous amount of domain-specific knowledge (and often does), it lacks wisdom. As such, the tool may be able to write code, but it won't be able to write code containing the nuances for specific or complex problems that require deep experience. Use ChatGPT to demo techniques, write small algorithms, and produce subroutines. You can even get ChatGPT to help you break down a bigger project into chunks, and then you can ask it to help you code those chunks. ... But you can do several things to help refine your code, debug problems, and anticipate errors that might crop up. My favorite new AI-enabled trick is to feed code to a different ChatGPT session (or a different chatbot entirely) and ask, "What's wrong with this code?"

How AI-enabled ‘bossware’ is being used to track and evaluate your work

Employee monitoring tools can increase efficiency with features such as facial recognition, predictive analytics, and real-time feedback for workers, allowing them to better prioritize tasks and even prevent burnout. When AI is added, the software can be used to track activity patterns, flag unusual behavior, and analyze communication for signs of stress or dissatisfaction, according to analysts and industry experts. It also generates productivity reports, classifies activities, and detects policy violations. ... LLMs are often used in predicting employee behaviors, including the risk of quitting, unionizing, or other actions, Moradi said. However, their role is mostly in analyzing personal communications, such as emails or messages. That can be tricky, because interpreting messages across different people can lead to incorrect inferences about someone’s job performance. “If an algorithm causes someone to be laid off, legal recourse for bias or other issues with the decision-making process is unclear, and it raises important questions about accountability in algorithmic decisions,” she said. The problem, Moradi explained, is that while AI can make bossware more efficient and insightful, the data being collected by LLMs is obfuscated. “So, knowing the way that these decisions [like layoffs] are made are obscured by these, like, black boxes,” Moradi said.

Attackers Can Manipulate AI Memory to Spread Lies

By crafting a series of seemingly innocuous prompts, an attacker can insert misleading data into an AI agent's memory bank, which the model later relies on to answer unrelated queries from other users. Researchers tested Minja on three AI agents developed on top of OpenAI's GPT-4 and GPT-4o models. These include RAP, a ReAct agent with retrieval-augmented generation that integrates past interactions into future decision-making for web shops; EHRAgent, a medical AI assistant designed to answer healthcare queries; and QA Agent, a custom-built question-answering model that reasons using Chain of Thought and is augmented by memory. A Minja attack on the EHRAgent caused the model to misattribute patient records, associating one patient's data with another. In the RAP web shop experiment, a Minja attack tricked the AI into recommending the wrong product, steering users searching for toothbrushes to a purchase page for floss picks. The QA Agent fell victim to manipulated memory prompts, producing incorrect answers to multiple-choice questions based on poisoned context. Minja operates in stages. An attacker interacts with an AI agent by submitting prompts that contain misleading contextual information. Referred to as indication prompts, they appear to be legitimate but contain subtle memory-altering instructions.

CISOs, are your medical devices secure? Attackers are watching closely

“To truly manage and prioritize risks, organizations need to look beyond technical scores and consider contextual risk factors that impact operations related to patient care. This can include identifying devices in critical care areas, legacy devices close to or past their end-of-life status, where any insecure communication protocols are, and how sensitive personal information is being stored,” Greenhalgh added. ... “For CISOs, the priority should be proactive engagement. First, implement real-time vulnerability tracking and ensure security patches can be deployed quickly without disrupting device functionality. Medical device security must be continuous—not just a checkpoint during development or regulatory submission. Second, regulatory alignment isn’t a one-time effort. The FDA now expects ongoing vulnerability monitoring, coordinated disclosure policies, and robust software patching strategies. Automating security processes—whether for SBOM (Software Bill of Materials) management, dependency tracking, or compliance reporting—reduces human error and improves response times. An SBOM is valuable not just for compliance but as a tool for tracking and mitigating vulnerabilities throughout a device’s lifecycle,” Ken Zalevsky, CEO of Vigilant Ops explained.

Can AI Teach You Empathy?

By leveraging AI-driven insights, banks can tailor their training programs to address specific skill gaps and enhance employee development. However, AI isn’t infallible, and it’s crucial for banks to implement tools that not only support learning but also foster a reliable and effective training environment. Striking the right balance between AI-driven training and human oversight ensures that these tools enhance employee growth without compromising accuracy or effectiveness. ... Experiential learning has long been a cornerstone of learning and development. Students, for example, who participate in experiential learning often develop a deeper understanding of the material and achieve statistically better outcomes than those who do not. While AI may not perfectly replicate a customer’s response, it provides new employees with a valuable opportunity to practice handling complex issues before interacting with real customers. AI-powered versions of these trainings can make it more accessible, allowing more employees to benefit. ... Many employees find it challenging to incorporate AI into their daily tasks and may need guidance to understand its value, especially in managing customer interactions. Some may also be resistant, fearing that AI could eventually replace their jobs, Huang says.

The Missing Piece in Platform Engineering: Recognizing Producers

The evolution of technology has shown us time and again that those who innovate are the ones who shape the future. Alan Kay’s words resonate strongly in the modern era, where software, artificial intelligence, and digital transformation continue to drive change across industries. ... “A Platform is a curated experience for engineers (the platform’s customers)” is a quote from the Team Topologies book. It is excellent and doesn’t contradict the platform business way of thinking, but it only calls out one side of the producer/consumer model. This is precisely the trap I fell into. When I worked with platform builders, we focused almost entirely on the application teams that consumed platform services. We rapidly became the blocker to those teams, just like the SRE and DevOps teams that came before us. We couldn’t onboard capabilities and features fast enough, meaning we were supporting the old ways while trying to build the new. ... Chris Plank, Enterprise Architect at NatWest, discusses this in our interview for his Platform Engineering Day talk: “We have since been set four challenges by leadership that I talk about: do things faster, do things simpler, enable inner sourcing, and deliver centralized capabilities in a self-service way… Our inner sourcing model will allow us to have multiple teams working on our platform… They are empowered to start contributing changes.”

Data Centers in Space: Separating Fact from Science Fiction

Among the many reasons for interest in orbital data centers is the potential for improved sustainability. However, the definition of a data center in space remains fluid, shaped by current technological limitations and evolving industry perspectives. Lonestar Data Holdings chairman and CEO Christopher Slott told Data Center Knowledge that his firm works from the definitions of a data center from industry standards bodies including the Uptime Institute and the Building Industry Consulting Service International (BICSI). ... Axiom Space plans to deploy larger ODC infrastructure in the coming years that are more similar to terrestrial data centers in terms of utility and capacity. The goal is to develop and operationalize terrestrial-grade cloud regions in low-Earth orbit (LEO). ... James noted that space presents the ultimate edge computing challenge – limited bandwidth, extreme conditions, and no room for failure. “To ensure resilience and autonomy, the platform incorporates automated rollbacks and self-healing capabilities through delta updates and health monitoring,” James said. ... With the Axiom Space deployment, the initial workloads will be small but scalable to the much larger ODC infrastructure that the company plans to deploy in the coming years. “Red Hat Device Edge enables secure, low-latency data processing directly on the ISS, allowing applications to run where the data is being generated,” James said.

CISA cybersecurity workforce faces cuts amid shifting US strategy

Analysts suggest these layoffs and funding cuts indicate a broader strategic shift in the U.S. government’s cybersecurity approach. Neil Shah, VP at Counterpoint Research, sees both risks and opportunities in the restructuring. “In the near to mid-term, this could weaken the US cybersecurity infrastructure. However, with AI proliferating, the US government likely has a Plan B — potentially shifting toward privatized cybersecurity infrastructure projects, similar to what we’re seeing with Project Stargate for AI,” Shah said. “If these gaps aren’t filled with viable alternatives, vulnerabilities could escalate from small-scale exploits to large-scale cyber incidents at state or federal levels. Signs point to a broader cybersecurity strategy reboot, with funding likely being redirected toward more efficient and sophisticated players rather than a purely vertical, government-led approach.” While some fear heightened risks, others argue the shift could lead to more tech-driven solutions. Faisal Kawoosa, founder and lead analyst at Techarc, views the move as part of a larger digital transformation. “Elon Musk’s role is not just about cost-cutting but also about leveraging technology to create more efficient systems,” Kawoosa said. “DOGE operates as a digital transformation program for US governance, exploring tech-first approaches to achieving similar or better results.”