Technology Risk Assessment

Three weeks ago, our Devsinc security architect, walked into my office with a chilling demonstration. Using quantum simulation software, she showed how RSA-2048 encryption – the same standard protecting billions of transactions daily – could theoretically be cracked in just 24 hours by a sufficiently powerful quantum computer. What took her classical computer billions of years to attempt, quantum algorithms could solve before tomorrow's sunrise. That moment crystallized a truth I've been grappling with: we're not just approaching a technological evolution; we're racing toward a cryptographic apocalypse. The quantum computing market tells a story of inevitable disruption, surging from $1.44 billion in 2025 to an expected $16.22 billion by 2034 – a staggering 30.88% CAGR that signals more than market enthusiasm. Research shows a 17-34% probability that cryptographically relevant quantum computers will exist by 2034, climbing to 79% by 2044. But here's what keeps me awake at night: adversaries are already employing "harvest now, decrypt later" strategies, collecting our encrypted data today to unlock tomorrow. For my fellow CTOs and CIOs: the U.S. National Security Memorandum 10 mandates full migration to post-quantum cryptography by 2035, with some agencies required to transition by 2030. This isn't optional. Ninety-five percent of cybersecurity experts rate quantum's threat to current systems as "very high," yet only 25% of organizations are actively addressing this in their risk management strategies. To the brilliant minds entering our industry: this represents the greatest cybersecurity challenge and opportunity of our generation. While quantum computing promises revolutionary advances in drug discovery, optimization, and AI, it simultaneously threatens the cryptographic foundation of our digital world. The demand for quantum-safe solutions will create entirely new career paths and industries. What moves me most is the democratizing potential of this challenge. Whether you're building solutions in Silicon Valley or Lahore, the quantum threat affects us all equally – and so does the opportunity to solve it. Post-quantum cryptography isn't just about surviving disruption; it's about architecting the secure digital infrastructure that will power humanity's next chapter. The countdown has begun. The question isn't whether quantum will break our current security – it's whether we'll be ready when it does.

AI agents are not yet safe for unsupervised use in enterprise environments The German Federal Office for Information Security (BSI) and France’s ANSSI have just released updated guidance on the secure integration of Large Language Models (LLMs). Their key message? Fully autonomous AI systems without human oversight are a security risk and should be avoided. As LLMs evolve into agentic systems capable of autonomous decision-making, the risks grow exponentially. From Prompt Injection attacks to unauthorized data access, the threats are real and increasingly sophisticated. The updated framework introduces Zero Trust principles tailored for LLMs: 1) No implicit trust: every interaction must be verified. 2) Strict authentication & least privilege access – even internal components must earn their permissions. 3) Continuous monitoring – not just outputs, but inputs must be validated and sanitized. 4) Sandboxing & session isolation – to prevent cross-session data leaks and persistent attacks. 5) Human-in-the-loop, i.e., critical decisions must remain under human control. Whether you're deploying chatbots, AI agents, or multimodal LLMs, this guidance is a must-read. It’s not just about compliance but about building trustworthy AI that respects privacy, integrity, and security. Bottom line: AI agents are not yet safe for unsupervised use in enterprise environments. If you're working with LLMs, it's time to rethink your architecture.

🚨 AI Privacy Risks & Mitigations Large Language Models (LLMs), by Isabel Barberá, is the 107-page report about AI & Privacy you were waiting for! [Bookmark & share below]. Topics covered: - Background "This section introduces Large Language Models, how they work, and their common applications. It also discusses performance evaluation measures, helping readers understand the foundational aspects of LLM systems." - Data Flow and Associated Privacy Risks in LLM Systems "Here, we explore how privacy risks emerge across different LLM service models, emphasizing the importance of understanding data flows throughout the AI lifecycle. This section also identifies risks and mitigations and examines roles and responsibilities under the AI Act and the GDPR." - Data Protection and Privacy Risk Assessment: Risk Identification "This section outlines criteria for identifying risks and provides examples of privacy risks specific to LLM systems. Developers and users can use this section as a starting point for identifying risks in their own systems." - Data Protection and Privacy Risk Assessment: Risk Estimation & Evaluation "Guidance on how to analyse, classify and assess privacy risks is provided here, with criteria for evaluating both the probability and severity of risks. This section explains how to derive a final risk evaluation to prioritize mitigation efforts effectively." - Data Protection and Privacy Risk Control "This section details risk treatment strategies, offering practical mitigation measures for common privacy risks in LLM systems. It also discusses residual risk acceptance and the iterative nature of risk management in AI systems." - Residual Risk Evaluation "Evaluating residual risks after mitigation is essential to ensure risks fall within acceptable thresholds and do not require further action. This section outlines how residual risks are evaluated to determine whether additional mitigation is needed or if the model or LLM system is ready for deployment." - Review & Monitor "This section covers the importance of reviewing risk management activities and maintaining a risk register. It also highlights the importance of continuous monitoring to detect emerging risks, assess real-world impact, and refine mitigation strategies." - Examples of LLM Systems’ Risk Assessments "Three detailed use cases are provided to demonstrate the application of the risk management framework in real-world scenarios. These examples illustrate how risks can be identified, assessed, and mitigated across various contexts." - Reference to Tools, Methodologies, Benchmarks, and Guidance "The final section compiles tools, evaluation metrics, benchmarks, methodologies, and standards to support developers and users in managing risks and evaluating the performance of LLM systems." 👉 Download it below. 👉 NEVER MISS my AI governance updates: join my newsletter's 58,500+ subscribers (below). #AI #AIGovernance #Privacy #DataProtection #AIRegulation #EDPB

Your dashboard says: ✅ 6212 tests passed ❌ 1230 failed ⏸️ 53 ignored Looks impressive, right? Actually... it tells you nothing. Too many teams still measure test coverage by the number of test cases executed. It feels like progress, but it’s often just busywork in disguise. Here’s a real-world story: A large Swiss bank had 1.3 million test cases. Not one of them found bugs. Why? Because an external vendor was paid per test case. More tests = more money. The result was just a mountain of useless tests. After digging in, they realized they only needed 60,000. That’s a 95% reduction. Wrong metric, wrong incentive, wrong outcome. Here’s a better way to measure coverage: Business Risk Coverage (BRC). Instead of counting test cases, ask: - Damage: What’s the impact if this feature fails? - Frequency: How often is it used in real workflows? - Risk = Damage × Frequency: Prioritize tests based on actual business risk. Now imagine your dashboard says: ✅ 35% Business Risk Covered ❌ 65% Not Covered Now that tells you something. Stop measuring quantity. Start measuring what matters. How are you measuring test coverage today? Curious to hear your take. → I’m on a mission to bring instant productivity to enterprise continuous testing. Follow me for insights from leaders around the industry and for my take on the future of #testautomation

The biggest threat to your data isn’t happening tomorrow. It happened yesterday. If you haven’t heard of HNDL (Harvest Now, Decrypt Later), your long-term data strategy has a massive blind spot. Here is the reality: State actors and cybercriminals are capturing your encrypted data today. They can’t read it yet, so they’re storing it in massive data vaults, waiting for the "Qday"—the moment quantum computers become powerful enough to break current encryption. If your data needs to stay private for 5, 10, or 20 years, it’s already at risk. What’s on the line? ↳ Intellectual Property (IP) and trade secrets. ↳ Government and identity data. ↳ Long-term financial records and contracts. ↳ Sensitive customer health data. How do we solve it? 🛠️ We cannot wait for quantum supremacy to react. The fix starts now: ↳ Inventory: Identify which data has a long shelf-life. ↳ Crypto-Agility: Move toward systems that can swap encryption methods without a total overhaul. ↳ Hybrid PQC: Implement Post-Quantum Cryptography alongside classical methods to ensure traffic captured today remains a mystery tomorrow. The transition to quantum-resistant security is a marathon, not a sprint. Are you tracking HNDL on your current risk register? Let’s discuss in the comments. 👇 P.S. If you want help mapping your exposure or building a PQC migration plan, drop me a message. ♻️ Share this post if it speaks to you, and follow me for more. #QuantumSecurity #PQC

🧠 Quantum computing: What business leaders need to do right now Right now, criminal and state-sponsored hackers are intercepting and storing encrypted data they cannot yet decode. Likely targets include everything from corporate secrets and medical records to legal agreements and military communications. Why would these actors bother to steal data they can’t read? Because they are betting on developments in quantum computing that will eventually let them crack this encrypted data wide open. This isn’t a fringe theory. The NSA (National Security Agency), NIST (National Institute of Standards and Technology), and ENISA (European Agency for Cybersecurity) are all treating this “harvest now, decrypt later” scenario as a live threat that is serious enough to demand immediate action. The NSA has mandated that all U.S. national security systems must transition to quantum-resistant cryptography by 2035—with new acquisitions required to be compliant by 2027. In Europe, ENISA issued updated guidance in April 2025 warning that the threat is “sufficient to warrant caution, and to warrant mitigating actions to be taken,” and recommending that organizations begin deploying post-quantum cryptography immediately. NIST has launched a parallel global effort to develop the new cryptographic standards on which these transitions will depend. The message from all three bodies is the same: Organizations run a grave risk if they wait to begin upgrades until quantum computers can break current encryption standards. That is the reason business leaders need to pay attention to quantum computing now — not because the technology is ready, but because the risk is grave, and the cost of preparation is trivial compared with the cost of being caught flat-footed. 🔗 Find out how in our new Fast Company article here: https://lnkd.in/g54y88UE.

Most quantum boardroom conversations end without an agenda. They end with a posture — "we're monitoring quantum developments," "we're taking it seriously". Neither statement produces a plan. The distinction matters because quantum creates three problem classes, each with a different urgency and a different cost of inaction. A generic posture misaddresses all three at once. The right response, for most leadership teams, has three parts. The first is to defend now. Post-quantum cryptography belongs on the enterprise risk agenda as a current priority. That means building visibility into cryptographic dependencies across the enterprise, identifying migration priorities, and mapping third-party exposure. This is the part of the quantum agenda that cannot wait. The second is to explore selectively. Most leadership teams do not need a wide portfolio of quantum pilots. They need a small number of focused efforts on high-value problems where the workload aligns with quantum's actual strengths — evaluated against the strongest available classical alternative. Each effort should be a targeted test: one specific problem, one clear classical benchmark, one honest evaluation. The third is to build options. For companies in simulation-relevant sectors — pharmaceuticals, advanced materials, energy — the right posture is modest investment in partnerships and early hardware collaborations. The goal is R&D workflows that are ready to integrate quantum subroutines when the technology matures. The companies that benefit most will not necessarily be those spending the most today. They will be the ones best positioned to move when the moment arrives. The most common failure on quantum is conflating the urgency of the three classes — treating all three as equally distant or equally immediate, when each has a different clock running. The organizations that get this right understand early which problem classes matter to their business, which ones to set aside, and what the distinction demands of them starting Monday morning. https://lnkd.in/gkymW7Xm

A distributed ledger built to last cannot rely on security assumptions that may not last. Quantum-safe preparation is necessary before the pressure arrives, because records and private keys must remain protected beyond today’s cryptography. The practical issue is time. Blockchain records can remain valuable for years, while the cryptographic methods that protect them may face new pressure as quantum computing advances. This does not mean organizations should panic or replace everything immediately. It means they need to understand which assets require long-term protection and where a migration path may be needed. Cryptographic signatures and private keys deserve special attention. If they protect ownership and access to assets, the question is not only whether they are safe today. The question is whether they can remain trusted across a longer horizon. Quantum-safe preparation is therefore a governance problem as much as a technical one. Organizations may need safer algorithms and migration planning, but the first step is knowing where the exposure sits. The goal is simple: preserve trust while security conditions change. For blockchain, durability is not enough if protection does not evolve with the system. #Blockchain #QuantumSecurity

The AI workflow produced great results, yet people did not feel safe relying on the output. ⛔ That was the situation I encountered in a client workshop in Brussels last week, and it is far more common than most organisations like to admit. The team had invested time and effort into designing an AI-supported workflow. The use case was clear, the technical setup was sound, the data quality was acceptable, and the people involved had already received training on how to use AI. Despite all of this, the workflow was barely used in practice. People ran the AI step, reviewed the output, and then quietly redid the work themselves. During the workshop, we mapped the real workflow together, step by step, focusing not on how the process was documented but on how the work actually happened on a normal working day. At one point, a participant looked at the whiteboard and said: “I only trust the result after I have checked it myself anyway.” That sentence shifted the entire conversation. As we continued mapping the process, a pattern became visible: Everyone validated AI outputs differently.  Some checked everything, even low-risk drafts.  Others barely checked high-risk decisions. Accountability was assumed but never explicitly defined. Human validation was happening constantly, but it was invisible, inconsistent, and highly personal. We redesigned the workflow and introduced a simple checklist for built-in human validation. 💡 This checklist replaced individual safety habits with a shared, explicit process. ✅ Define the risk level of the output. Clarify whether the AI output is a draft, a recommendation, or a decision with external impact. ✅ Decide if validation is required. Make it explicit which outputs require human review and which can flow through without intervention. ✅ Specify the validation moment. Define when validation happens in the workflow and before which downstream step. ✅ Assign clear responsibility. Name the role that validates the output and the role that makes the final decision. ✅ Separate generation from judgment. Ensure the AI prepares content or options, while humans remain accountable for approval and outcomes. ✅ Remove unnecessary checks. Regularly review the workflow to eliminate validation steps that add friction without reducing risk. Once this checklist was applied, people felt much more confident about the AI output because they knew when human judgment was required. 👉 Is human validation in your AI workflows clearly designed, or is it still improvised? Let’s discuss.

𝗜𝗧 𝘃𝘀 𝗢𝗧 𝘃𝘂𝗹𝗻𝗲𝗿𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝗺𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁 Same goal. Different constraints, timelines and risk decisions. A vulnerability-management process that works well in enterprise IT cannot simply be copied into a live OT environment unchanged. Traditional IT scanning methods can potentially disrupt OT components, particularly older or fragile systems. OT guidance therefore recommends using vulnerability scanning only where it is safe and feasible, ideally after testing on representative systems or with vendor and operations approval. ▪️ 𝗜𝗻 𝗜𝗧 The usual approach may include: ✓ Broader active scanning ✓ Regular patch cycles ✓ More flexible reboot scheduling ✓ Prioritisation based on exposure, exploitability, business impact and CVSS Even in IT, CVSS should not be the only decision factor. ▪️ 𝗜𝗻 𝗢𝗧 The same actions require more operational context: ✓ Use passive discovery or carefully approved active scanning ✓ Test patches before deployment ✓ Coordinate reboots with operations ✓ Align changes with maintenance windows ✓ Confirm vendor support and firmware compatibility ✓ Evaluate safety, availability and process consequences NIST SP 800-82 Rev. 3 specifically frames OT security around its unique performance, reliability and safety requirements. ISA/IEC 62443-2-3 provides a structured approach to patch identification, testing, deployment, verification and supplier coordination in IACS environments. 𝗣𝗿𝗶𝗼𝗿𝗶𝘁𝗶𝘀𝗮𝘁𝗶𝗼𝗻 𝗺𝘂𝘀𝘁 𝗯𝗲 𝗿𝗶𝘀𝗸-𝗯𝗮𝘀𝗲𝗱 In OT, a high CVSS score does not automatically mean “patch immediately.” Also consider: ✓ Is the asset exposed? ✓ Is the vulnerability known to be exploited? ✓ What process does the asset control? ✓ Could exploitation affect safety or production? ✓ Is a tested patch available? ✓ Can compensating controls reduce the risk until the next window? The CISA Known Exploited Vulnerabilities catalogue can help identify vulnerabilities with evidence of active exploitation, but the final treatment decision still needs asset and process context. 𝗤𝘂𝗶𝗰𝗸 𝗿𝘂𝗹𝗲 𝗼𝗳 𝘁𝗵𝘂𝗺𝗯 In IT, the vulnerability often drives the timeline. In OT, the vulnerability determines the urgency, while operations, safety and the production schedule determine how the risk is treated. That may mean: 𝗣𝗮𝘁𝗰𝗵𝗶𝗻𝗴 or 𝗦𝗲𝗴𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 + 𝗮𝗹𝗹𝗼𝘄𝗹𝗶𝘀𝘁𝗶𝗻𝗴 + 𝗺𝗼𝗻𝗶𝘁𝗼𝗿𝗶𝗻𝗴 + 𝗮 𝗽𝗹𝗮𝗻𝗻𝗲𝗱 𝗽𝗮𝘁𝗰𝗵 𝘄𝗶𝗻𝗱𝗼𝘄 The objective is not to delay remediation. It is to reduce cyber risk without introducing an operational or safety incident through the remediation itself. #OTSecurity #VulnerabilityManagement #PatchManagement #IndustrialCybersecurity #ICSSecurity #IEC62443