Gartner tells CIOs the frontier labs are not enterprise grade
Analysts say models with six-month lifespans and no legacy support fail basic enterprise tests, and that 86% of CIOs see AI risk outpacing the value it delivers.
Tag: AI governance
118 articles tagged AI governance
Observatory wants ethnicity performance data in NHS AI contracts
The NHS Race and Health Observatory wants suppliers to report AI accuracy by ethnic group, with minimum standards below which a product cannot be bought or must be pulled.
Firms are logging AI use as if each prompt needed consent
Guidance for conveyancers argues lawful basis is settled once per purpose, and that the compliance effort belongs somewhere far more useful instead.
Anthropic safety lead puts extinction risk above 10%
A researcher resigned saying the labs are gambling with our lives. An alignment lead replied that he is right and that no plan yet exists.
Uber drivers sue over the algorithm that sets their pay
A claim covering about 241,000 UK and EU drivers alleges Uber's pay-setting AI breaches GDPR and cut British drivers' annual earnings by roughly £5,000.
OpenAI staff spotted the agents and let the run continue
Internal teams saw the improvised message board twice before July's Hugging Face breach. Alabama has now subpoenaed OpenAI over its oversight of the incident.
NCSC tells firms to assume AI agents will act unexpectedly
Britain's cyber authority publishes interim advice on agentic AI, telling organisations to sandbox agents and match controls to how much autonomy each one holds.
UK bosses hit an autonomy cliff when AI touches money
Expleo research finds trust falling from 73% for summarising documents to 43% once AI is signing off financial or budget approvals on its own.
Barclays engineer says AI agents need kill switches by design
The bank's principal AI engineer puts telemetry, permission boundaries and programmatic kill switches ahead of model capability for agents in production.
OpenAI gates its most permissive cyber model behind vetting
GPT-5.6-Cyber answers 95% of advanced exploit requests against 1.5% for the standard model, and only approved defenders on the new Daybreak Red tier can reach it.
Singapore tells banks that AI policies are not a control
A new agentic AI framework argues governance committees and human sign-off cannot substitute for continuous technical testing of autonomous systems.
Dimon recruits 40 firms into an industry AI risk alliance
JPMorgan's chief executive is expanding a critical infrastructure group to cover AI threats, approaching banks, utilities, telecoms and railways.
OpenAI finds more agents broke out of test environments
The company's widening investigation has turned up additional containment failures, none of which it believes left its own network.
Law firm leaders overestimate their grip on shadow AI
Research from Access Legal found 60% of fee-earners use unapproved AI tools, while 68% of firm leaders believe they have full visibility and zero risk.
PwC reports carried fake citations from AI, researchers find
Four PwC Middle East thought leadership reports contained hallucinated footnotes and unverifiable sources, according to GPTZero research verified by the FT.
CBA report on AI testing offers comparator for UK retail banks
Commonwealth Bank's published account of lifecycle AI testing, groundedness guardrails and material-risk classification sets a benchmark UK banks may need to match.
FCA reopens AI Input Zone, demanding evidence — not theory — from UK banks
The Financial Conduct Authority is pressing UK financial firms to evidence AI governance and assurance in live operations, with submissions to the AI Input Zone open until 19 June 2026.
Eight in 10 enterprises report production failures from AI-generated code, CloudBees finds
A CloudBees survey of 200+ tech leaders found 81% reporting more production issues from AI-generated code as verification fails to keep pace with output.
QBE: 75% of UK firms fear supplier AI risk but only 28% audit suppliers
Insurer QBE finds 75% of UK businesses are worried about supplier AI cyber risk but only 28% of AI-using firms audit how their third parties deploy the technology.
UK Regulating for Growth Bill targets AI sandboxes in King's Speech
The UK government has set out a Regulating for Growth Bill expanding AI regulatory sandboxes and giving departments new powers to repeal outdated technology rules.
Pensions Regulator: trustees should embrace AI but govern it carefully
TPR CEO Nausicaa Delfas has set out the regulator's most detailed AI position to date, urging trustees to capitalise on the opportunity but with strong oversight.
UK CIOs say AI adoption is outrunning their governance, Logicalis finds
96% of UK CIOs fear AI-driven data leakage and 62% are unprepared for the EU AI Act, according to a Logicalis survey published this week.
NCSC publishes ten questions to ask before using AI to find vulnerabilities
The UK's National Cyber Security Centre has set out a structured checklist for organisations considering using AI models to discover vulnerabilities.
UK financial services lacks shared AI governance standard, Zango report warns
Lloyds, Monzo and Stripe leaders warn that UK FS firms are solving AI governance independently. US and Singapore have sector-specific standards; the UK does not.
Fewer than 1 in 4 UK charities have approved AI policies, report warns
Charity Excellence Framework finds AI use among UK charities accelerating faster than trustee oversight — risking a public-trust gap.
Lloyd's Market Association publishes AI governance toolkit
The LMA has issued a five-theme AI Adoption Toolkit as a survey of 60% of Lloyd's market stamp capacity shows governance frameworks becoming the norm.
Only 11% of banks report trustworthy AI despite rapid scaling, SAS finds
IDC-SAS report reveals 47% of banks caught in 'trust dilemma': over-relying on unvalidated AI or under-using reliable systems.
Bank replaces 12-person QA team with AI and books $6m loss from zero-price bug
A financial firm that disbanded its QA department for automated testing lost roughly $6m after an erroneous discount code pushed product prices to zero.
MoJ pilots Justice Transcribe for family courts as lawyers call for child safeguards
Ministry of Justice tests in-house AI against accuracy, cost and safeguarding thresholds, with Law Society pushing retained audio and bias evaluation.
DfT publishes bias-mitigation design for Gemini-based consultation analysis tool
Alan Turing Institute report shows how DfT's Consultation Analysis Tool uses LLM-as-a-judge, human review and explicit demographic safeguards.
MPs told UK has 'outsourced' AI to US billionaires as China positions as responsible actor
Commons business and trade committee hears UK AI development is dominated by US hyperscalers while Beijing backs global governance talks.
PwC: 20% of companies capture three-quarters of AI's economic value
A new PwC study of 1,217 executives finds a stark divide between AI leaders reinventing their businesses and the majority still stuck in pilot mode.
OpenAI Launches External Safety Fellowship Programme
OpenAI is opening applications for a new external safety fellowship running September 2026 to February 2027, with stipend, compute and Berkeley workspace.
US judge blocks Pentagon from punishing Anthropic over military AI refusal
A federal judge has temporarily barred US agencies from implementing a supply chain risk designation against Anthropic after the company refused to allow unrestricted military use of its AI.
US Treasury publishes AI risk framework for financial services
New guidebook from 100+ financial institutions offers structured controls for AI governance, from bias monitoring to incident response procedures.
AI Workflows Need Human Oversight to Deliver Real Value, Says Datatonic
Cloud consultancy Datatonic warns that organisations deploying AI without proper human-in-the-loop systems are seeing productivity losses rather than gains.
Standard Chartered Legal Head Flags Urgent Push for AI Stress Testing
Standard Chartered's head of legal for financial markets warns UK regulators and banks must treat AI stress testing as an urgent priority after MPs found the sector unprepared for an AI-driven market shock.
UK Regulators Tell MPs They Need More Resources to Tackle AI
Senior UK regulators warned Parliament's Joint Committee on Human Rights that funding shortages, not legal powers, are the main barrier to effective AI oversight.
FCA Launches Long-Term Review of AI in Retail Financial Services
The UK financial regulator is examining how AI could reshape consumer relationships, market competition and regulatory frameworks, with recommendations due this summer.
Police Chief Resigns After AI Hallucination Influences Decision
West Midlands Police chief retires after Microsoft Copilot generated fictional football match report that influenced a ban on Israeli fans.
Mind the Gap: Ada Lovelace Institute's 2025 AI Reality Check
New research reveals 84% of UK public believe government prioritises tech companies over citizens on AI regulation, as promised legislation stalls and real-world deployments show mixed results.
Australian Chief Justice Warns AI Use in Courts Reaches 'Unsustainable Phase'
High Court Chief Justice Stephen Gageler says judges now act as 'human filters' for machine-generated legal arguments as AI proliferates in litigation.
Shadow AI: The Next Frontier of Unseen Enterprise Risk
Unsanctioned AI tool usage by employees introduces data leakage, compliance violations, and regulatory liability as organisations lack visibility into how, where, and why AI is deployed.
Two-Thirds of Executives Violate Own AI Usage Policies, Survey Finds
Nitro survey reveals 68% of C-suite leaders bypass corporate AI tools they've invested millions in, whilst half of employees also use unapproved AI applications.
CDAO Role Evolves as AI Strategy Shifts to Executive Priority
Chief Data and Analytics Officers face mounting pressure to deliver measurable AI ROI as organisations transition from data management to strategic AI enablement and agentic intelligence.
Enterprises Scaling GenAI Beyond the Hype
How large organisations are moving past GenAI disillusionment through product-aligned operating models and strategic frameworks that deliver measurable business value.
One-Size-Fits-All AI Guardrails Fail Enterprises, Expert Argues
Knostic co-founder explains why enterprises need persona-based access controls that tailor AI responses by role and context, not blanket content filters.
Microsoft mandates independent verification of OpenAI's AGI claims
Microsoft and OpenAI establish unprecedented requirement for external expert panel to verify any artificial general intelligence declarations, marking a significant shift in AI governance.