Daily Tech Digest - February 28, 2024

3 guiding principles of data security in the AI era

Securing the AI: All AI deployments – including data, pipelines, and model output – cannot be secured in isolation. Security programs need to account for the context in which AI systems are used and their impact on sensitive data exposure, effective access, and regulatory compliance. Securing the AI model itself means identifying model risks, over-permissive access, and data flow violations throughout the AI pipeline. Securing from AI: Just like most new technologies, artificial intelligence is a double-edged sword. Cyber criminals are increasingly turning to AI to generate and execute attacks at scale. Attackers are currently leveraging generative AI to create malicious software, draft convincing phishing emails, and spread disinformation online via deep fakes. There’s also the possibility that attackers could compromise generative AI tools and large language models themselves. ... Securing with AI: How can AI become an integral part of your defense strategy? Embracing the technology for defense opens possibilities for defenders to anticipate, track, and thwart cyberattacks to an unprecedented degree. AI offers a streamlined way to sift through threats and prioritize which ones are most critical, saving security analysts countless hours. 


Web3 messaging: Fostering a new era of privacy and interoperability

Designed to be interoperable with various decentralized applications (DApps) and blockchain networks, Web3 messaging protocols enable developers to seamlessly integrate messaging functionality into their decentralized services — a stark contrast to their traditional equivalents that host closed ecosystems, which limit communication with users on other platforms. Beoble, a communication infrastructure and ecosystem that allows users to chat between wallets, is one of the Web3 messaging platforms ready to change how people use digital communication. The platform comprises a web-based chat application and a toolkit for seamless integration with DApps. Dubbed “WhatsApp for Web3,” Beoble removes the need for login methods like Twitter or Discord, instead mandating only a wallet for access. Users can log in using their wallets and send texts, images, videos, links and files across blockchain networks. Blockchain app users can utilize emojis and nonfungible token (NFT) stickers in their digital communication with Beoble, adding a layer of personality to their conversations. 


As data takes center stage, Codified wants to bring flexibility to governance

As Gupta sees it, many large companies are authoring policies and trying to implement them in various ways, but he sees software that is too rigid for today’s use cases, leaving them vulnerable, especially when they have to change policy. He wants to change that by translating policy into code that can be implemented in a variety of ways, connected to various applications that need access to the data, and easily changed when new customers or user categories come along. “We let you author policies in natural language, in a declarative way or using a UI - pick your favorite way - but when those policies are authored, we can codify them into something that can be implemented in a number of ways and can be very easily changed,” he said. To that end, the company also enables customers to set conditions, such as whether you’ve had security training in the last 365 days, or you’re already part of a team working on a sensitive project. Ultimately, this enables companies to set hard-coded data access rules based on who the employee is and the applications they are using or projects they are part of, rather than relying on creating groups on which to base these rules.


Looking Forward, Looking Back: A Quarter Century as a CISO

The first distributed denial of service (DDoS) attack occurred in 1999, followed by Code Red and Nimda worm cyberattacks that targeted web servers in 2001, and SQL Slammer in 2003 which spread rapidly and brought focus on the need to patch vulnerable systems. The end of the millennium also brought Y2K and the Millennium Bug, which exposed the vulnerability of existing computing infrastructures that formatted dates with only the final two digits and raised the profile of CISOs and other security professionals. Organizations recognized the necessity of dedicated executives responsible for managing cybersecurity risks. ... CISOs were soon making the news, and not always in a good way. Former Uber CISO Joe Sullivan was found guilty of felony obstruction of justice and concealing a data breach in October 2022. The following month, CISO Lea Kissner of Twitter (now X) resigned along with the company’s chief privacy officer and its chief compliance officer over concerns that Twitter’s new leadership was pushing for the release of products and platform changes without effective security reviews.


How Generative AI is Revamping Digital Transformation to Change How Businesses Scale

Crucially, generative AI can help to tailor the dining experience for customers in a way that significantly improves the quality of in-house or takeaway eating. This is achieved by GenAI models analyzing data like guest preferences, dietary restrictions, past orders, and behavior to offer personalized menu items and even recommend food and drink pairings. Generative AI will even be capable of using available datasets to generate offers on the fly as an instant call-to-action (CTA) if it deems an online visitor isn't yet ready to convert their interest into action. We're already seeing leading global restaurants announce the implementation of generative AI for their processes. ... Generative AI became the technological buzzword of 2023, and for good reason. However, there will be many hurdles to overcome in the development of the technology before it drives widespread digital transformation. Regulatory hurdles may be tricky to overcome due to issues in how AI programs can handle private data and utilize intellectual property (IP). Quality shortcomings could also cause issues in governance among early LLMs, and we've seen plenty of cases where language models "hallucinate" when dealing with unusual queries.


NIST CSF 2.0 released, to help all organizations, not just those in critical infrastructure

The CSF’s governance component emphasizes that cybersecurity is a major source of enterprise risk that senior leaders should consider alongside others, such as finance and reputation. “Developed by working closely with stakeholders and reflecting the most recent cybersecurity challenges and management practices, this update aims to make the framework even more relevant to a wider swath of users in the United States and abroad,” according to Kevin Stine, chief of NIST’s Applied Cybersecurity Division. ... The framework’s core is now organized around six key functions: Identify, Protect, Detect, Respond, and Recover, along with CSF 2.0’s newly added Govern function. When considered together, these functions provide a comprehensive view of the life cycle for managing cybersecurity risk. The updated framework anticipates that organizations will come to the CSF with varying needs and degrees of experience implementing cybersecurity tools. New adopters can learn from other users’ successes and select their topic of interest from a new set of implementation examples and quick-start guides designed for specific types of users...


Even LLMs need education—quality data makes LLMs overperform

Like any student, LLMs need a good source text to produce good outputs. As Satish Jayanthi of CTO and co-founder of Coalesce told us, “If there were LLMs in the 1700s, and we asked ChatGPT back then whether the earth is round or flat and ChatGPT said it was flat, that would be because that's what we fed it to believe as the truth. What we give and share with an LLM and how we train it will influence the output.” Organizations that operate in specialized domains will likely need to train or fine-tune LLMs of specialized data that teaches those models how to understand that domain. Here at Stack Overflow, we’re working with our Teams customers to incorporate their internal data into GenAI systems. When Intuit was ramping up their GenAI program, they knew that they needed to train their own LLMs to work effectively in financial domains that use tons of specialized language. And IBM, in creating an enterprise-ready GenAI platform in watsonx, made sure to create multiple domain-aware models for code, geospatial data, IT events, and molecules.


State of FinOps 2024: Reducing Waste and Embracing AI

Engineers remain the biggest beneficiary of FinOps observability, even though "engineering enablement" has dropped to a lower position in the report's ranking of surveyed priorities. This indicates that engineers are those best suited to responding to a sudden change in cost metrics. The report observes that the "engineering persona" is reported as getting the most value from both "FinOps training and self-service reporting." ... While waste reduction is a common driver across all respondents, segmenting the survey by cloud spend revealed that those with smaller budgets would tend to then prioritise improvements in the accuracy of billing forecasts. The report states that these respondents faced the challenge of understanding "the trajectory of spending" prior to it "getting out of hand." Most invested in low-effort solutions such as "manual adjustments" to generated forecast data. In contrast, those with larger budgets tended to prioritise the optimisation of commitment-based discounts to benefit from economies of scale. This included the right-sizing of "reserved instances, savings plans, committed use discounts," as well as specific negotiated discounts.


How to Develop an Effective Governance Risk and Compliance Strategy

“Overcoming silos and fostering communication needs to begin at the top,” Rothaar, says in an email interview. Furthermore, aligning GRC goals with broader business objectives ensures both executive management and individual departments recognize the impact that GRC initiatives have on organizational success. “Promoting a culture of communication with open dialogue and knowledge-sharing is essential to a successful and efficient GRC strategy,” she says. Ringel says organizations need to promote awareness and engagement with risk and compliance, because they influence every member of the organization. “You are only as strong as your weakest link when it comes to risk, so making sure everyone is on the same page and treating risk and compliance smartly is key,” she explains. Compliance is less directly obvious, but if those values are not communicated through every department--product design, development, customer support, marketing, and sales -- the end product will reflect that disconnect. “Not every employee needs to know specific regulations, but everyone needs to share the values of data governance and compliance,” Ringel says.


Data storage problems and how to fix them

When undertaking the journey to digitisation, it’s important to consider the issues and challenges and more importantly – know how to avoid them. ... It’s wise not to attempt a massive data overhaul all at once, especially before you’ve considered what data is valuable, how and where you will store the data and investigated the different options and models available. It all depends on the scope of transformation and the state the organisation is in. For start-ups, it’s a green field and the experience is as good as the plan and its periodic inspection and adaptation. For organisations with historic data to migrate, it can get complex. I have experienced both and the key was to have identified what data is valuable, a clear cut off date and policy on how far back we digitise. ... If you are unsure on where to start, consult an expert to determine the best solutions and view the initial costs as an investment. Digital transformation of data brings the benefits of creating efficiency and timesaving and with those, reduced costs. The long-term benefit can far outweigh the upfront costs. Digital systems are typically faster and more efficient than manual systems. 



Quote for the day:

"Nothing is so potent as the silent influence of a good example." -- James Kent

No comments:

Post a Comment