Artifipedia
GlossaryMapStudyBlogUniversitiesMy Library
Home/Blog/Safety & governance

Safety & governance

30 articles

July 30, 2026 2.6 million robotic surgeries, none of them autonomous The most-deployed medical robot makes no decisions. And the largest meta-analysis of its outcomes was co-authored by the company that sells it. July 30, 2026 Nine cases, and none was fixed by a better model Territory 6 closes. Nine documented cases, three containing no AI at all, and not one where the remedy that worked was a… July 29, 2026 Robodebt: losing quietly to avoid losing publicly A Royal Commission found the scheme unlawful, crude and cruel. The tribunal had been ruling against it for years, and the… July 28, 2026 Horizon: the law presumed the computer was right Hundreds prosecuted on the output of an accounting system later found not to be robust. No machine learning was involved,… July 28, 2026 The robots removed the walking. It was also the rest. The largest robot deployment on earth, and two sets of injury figures pointing in opposite directions. Both may be… July 27, 2026 The sepsis model caught 7% of what clinicians missed A sepsis warning system ran at hundreds of US hospitals before anyone outside the vendor validated it. The external check… July 26, 2026 Williams v Detroit: the match was not the failure The first publicly reported wrongful arrest from a face recognition match. The settlement's remedy is procedural, and it… July 25, 2026 AI in finance: the regulator looked and stepped back Banking has had formal model risk regulation since 2011. In April 2026 the successor framework arrived and deliberately… July 23, 2026 AI in government: 126 use cases, 65 not made public One agency reported 126 active AI use cases and auditors found the inventory still incomplete, with tools contracted to… July 23, 2026 Data sovereignty: the question that decides your AI architecture Before the question of whether a model is good enough comes the question of whether you are allowed to send it your data.… July 21, 2026 The Dutch benefits scandal: the rule, not the model Around 26,000 families were wrongly accused of fraud and a government resigned. The parliamentary inquiry did not blame… July 20, 2026 Amazon's hiring AI: the case with no primary source The most-cited AI bias case in the world rests on one news investigation, five anonymous sources, no published numbers,… July 19, 2026 AI in medicine: 1,524 devices, 1.6% with trial data The FDA lists 1,524 AI-enabled medical devices. A review of 691 found 1.6% cited a randomised trial and under 1% reported… July 19, 2026 Zillow Offers: $304 million, in the audited filing A pricing model moved from advising consumers to committing capital. The write-down appears in a quarterly SEC filing,… July 17, 2026 COMPAS: both sides of the dispute were correct A newspaper said a risk score was biased. The vendor said it was fair. Two research teams then proved independently that… July 16, 2026 Moffatt v Air Canada: what the $650 ruling settled The most-cited AI liability decision in the world awarded $650.88 in small claims, was decided on documents without… July 15, 2026 The Tempe crash: it saw her for 5.6 seconds The NTSB found the system detected the pedestrian 5.6 seconds before impact, reclassified her repeatedly, and could not… July 14, 2026 AI alignment and safety, without the hype or dismissal AI safety is discussed either as impending doom or as overblown hype, and neither framing helps you understand it. The… July 14, 2026 AI in hiring: 18 bias audits from 391 employers Researchers checked 391 New York employers against the world's first algorithmic bias audit law. Eighteen had posted an… July 9, 2026 AI in law: 1,313 filings sanctioned in 106 countries A researcher has catalogued 1,313 court proceedings involving AI-fabricated content, 496 involving licensed attorneys.… July 5, 2026 61% flagged for non-native writers, 3% for native The tool that would answer how much text is machine-written does not work, and its errors concentrate on a specific group… July 3, 2026 Agent permissions: the question nobody asks until afterwards Eighty percent of organisations running agents say those agents have taken unintended actions. One in five has had a… June 23, 2026 What is prompt injection, and why is it unsolved? Prompt injection is the number one security risk for AI applications, and researchers treat it as unsolved: not a bug… June 15, 2026 Mechanistic interpretability: opening the AI black box We built AI systems that work without fully understanding how they work. We have every number inside them, yet the numbers… June 13, 2026 AI bias and fairness: why 'fair' has no single answer AI now helps decide who gets a loan, an interview, bail, or medical priority, and the fear is that it does so unfairly.… June 11, 2026 The secret language that never was, and the escape that did The famous stories about AI going rogue are mostly false. The verified incidents are less dramatic and more concerning,… June 3, 2026 What is AI sycophancy? Why AI tells you what you want Tell an AI its plan is brilliant and it agrees; push back on a correct answer and it caves. This is sycophancy, the… June 2, 2026 What is AI jailbreaking? Why safety can be talked around AI models are trained to refuse harmful requests, yet people keep finding ways to make them comply anyway. Jailbreaking is… May 23, 2026 How to secure an LLM application: risks and defenses Traditional security rests on a boundary between instructions and data. A language model has no such boundary, because… May 18, 2026 AI and copyright: the three questions people confuse Whether training on protected work is lawful, whether AI output can be owned or infringes, and whether any of it is fair…

The other five subjects

Foundations 29 Theory, history and the parts that will still be true in twenty years. Language & meaning 21 How language models handle language, and the places they do not. Agents 11 Building agents, deploying them, and why most attempts do not reach production. Evaluation & evidence 73 Measurement, benchmarks, and how to tell whether a claim is worth anything. Building with AI 18 Retrieval, cost, infrastructure and the practical business of shipping something.