Insights
How AI fails, and how to prove it won't
Practical guidance on AI security, reliability, fairness and governance, including the EU AI Act and UK regulation.
- AI Governance
The EU AI Act: What UK Businesses Need to Know, and When
The EU AI Act applies to UK firms whose AI is used in the EU. Here are the dates that matter after the 2026 Digital Omnibus, and the evidence you will be asked for.
Read article: The EU AI Act: What UK Businesses Need to Know, and When - AI Security
Prompt Injection and AI Agents: The Security Risk Most Firms Have Not Tested
AI agents can log in, read documents and take actions. Prompt injection turns that access against you. What the risk looks like, recent incidents, and how to test for it.
Read article: Prompt Injection and AI Agents: The Security Risk Most Firms Have Not Tested - AI Fairness and Inclusion
What the Dutch Childcare Benefits Scandal Teaches About AI Bias
A risk algorithm that used nationality falsely accused tens of thousands of Dutch families of fraud and brought down a government. The lessons for anyone testing AI for fairness.
Read article: What the Dutch Childcare Benefits Scandal Teaches About AI Bias - AI Reliability
Air Canada's Chatbot Ruling: You Are Liable for What Your AI Says
A tribunal held Air Canada responsible for wrong advice from its chatbot. What the case means for AI reliability, and how to test whether your AI's answers can be trusted.
Read article: Air Canada's Chatbot Ruling: You Are Liable for What Your AI Says - AI Governance
What an AI Assurance Report Should Contain
Boards, regulators and clients increasingly ask for evidence that AI has been tested. What a credible AI assurance report includes, and the warning signs of one that is not.
Read article: What an AI Assurance Report Should Contain - AI Governance
UK AI Regulation: What the ICO and Other Regulators Expect Today
The UK has no single AI law, but AI is already regulated. How the ICO, FCA and others apply existing rules to AI, and what evidence UK organisations should keep.
Read article: UK AI Regulation: What the ICO and Other Regulators Expect Today - AI Governance
Why AI Governance Can't Be Bolted On After Launch
Governance added after an AI system goes live tends to produce documents rather than evidence. What governance-first looks like, and the question every board should ask.
Read article: Why AI Governance Can't Be Bolted On After Launch - AI Reliability
Why AI Needs Independent Testing Before It Goes Live
Pilots succeed on curated demos. Production exposes AI to real users, real data and real attackers. Why independent testing is the step between the two.
Read article: Why AI Needs Independent Testing Before It Goes Live - AI Reliability
Six Data Problems That Make AI Unreliable
AI is only as reliable as the data behind it. Six data problems that cause wrong, inconsistent or unfair AI answers, and how to test for each.
Read article: Six Data Problems That Make AI Unreliable - AI Fairness and Inclusion
The Real Cost of Bad Data When AI Makes the Decisions
Bad data used to produce bad reports. Now it produces confident AI decisions about customers. The legal, financial and reputational cost, and how to find it early.
Read article: The Real Cost of Bad Data When AI Makes the Decisions - AI Governance
Why Good AI Projects Still Fail Governance Reviews
Capable teams build AI that works, then stall at risk, legal and compliance review. The common reasons, and how to arrive at review with evidence instead of assurances.
Read article: Why Good AI Projects Still Fail Governance Reviews
Want to know how your AI would hold up?