Frontier Models Face Gatekeeping

Axios reported that the US government is expanding frontier AI testing with Google DeepMind, Microsoft and xAI. The coverage describes a move toward pre-deployment checks for security-sensitive AI capabilities before public release. Reuters reported that major publishers filed a proposed class action against Meta over alleged use of books and scientific works to train Llama. The claim adds another large rights-holder dispute to the unresolved fair-use fight over AI training data.

Assurance hardens, healthcare sandboxes scale, and exams draw clearer AI lines

The Digital Assurance Playbook took effect from 1 April and requires government assurance teams to test AI initiatives against the UK government’s AI Playbook and the Algorithmic Transparency Reporting Standard. The MHRA expanded its AI Airlock programme with three years of funding to scale sandbox testing for AI as a medical device. The FCA announced the second cohort for AI Live Testing, extending supervised live experimentation in financial services and committing to further good and poor practice output later this year. The PRA Business Plan 2026/27 made responsible AI adoption and monitoring of firms’ AI use an explicit supervisory priority for the year ahead. Ofqual published Artificial intelligence malpractice and assessment - advice note, clarifying how existing qualification rules apply to AI-related misuse.

AI Guardrails Reach Children and Finance

Associated Press reported that Roblox will require facial scans for users under 16 in Indonesia under new rules for minors on digital platforms. The rollout creates age-based accounts and restricts chat access for children who have not completed facial age estimation. Insurance Journal reported Grant Thornton findings that insurers are gaining from AI but still face governance gaps. The survey found governance or compliance weaknesses linked to failed or underperforming AI projects, with limited confidence in independent review readiness. ABA Journal reported that Harvey will expand its legal AI presence with a new Chicago office. The move reflects continuing institutional investment in legal AI tools despite heightened scrutiny of accuracy, workflow control and professional responsibility.
Subscribe to