Anthropic and Accenture each expect to invest at least $1 billion over five years in embedded evaluation led by Accenture's Faculty business. The planned work includes model evaluation, red-teaming, alignment assessment and safeguard testing, with evaluator access intended to resemble employee access.
Why it matters: A standing outside team with deep access could make safety commitments, incidents and release decisions more observable than occasional external tests. The funding relationship, access rules and publication rights will determine whether the arrangement…
Researchers from Google and Google DeepMind launched the DeepMind Institute, a public forum for work on artificial general intelligence and society. Its first essays cover reasoning transparency, economic policy, global access, human flourishing and capability-based testing of frontier models.
Why it matters: The institute makes policy and technical-safety proposals from researchers close to a frontier-model lab publicly inspectable and open to wider debate. Its reasoning-transparency work is especially relevant to teams deciding whether internal model traces…
European Commission President Ursula von der Leyen said the Commission would invite leading frontier AI laboratories to discuss industry efforts to pace frontier development. She also proposed work with Canada, the United Kingdom and other partners on model evaluation, verification, early warning and AI security.
Why it matters: A senior EU policy signal links voluntary laboratory behavior with international testing and warning mechanisms. If converted into concrete commitments, it could affect release gates, evidence duties and incident reporting for frontier models.
The UK Parliament's Joint Committee on Human Rights published a 100-page report recommending a cross-sector AI Bill with risk-tiered duties across the full AI lifecycle. It calls for mandatory transparency and differentiated due diligence, prohibitions on uses incompatible with human rights, a single oversight body, and compulsory pre-release AISI testing of powerful models with authority to advise that unsafe systems be blocked or withdrawn.
Why it matters: The proposal would move responsibility beyond deployers to developers, providers and other supply-chain actors, while turning model access, technical disclosure, testing and market withdrawal into statutory governance questions. It is one of the clearest…
Anthropic CEO Dario Amodei committed the company to inviting a permanent external review team with employee-like access to relevant risk-assessment workspaces, tools, permissions and staff. The proposed contract would let reviewers publish key findings about risks, incidents, practices and denied access without Anthropic editorial control, subject to narrow security, legal and third-party confidentiality redactions.
Why it matters: The commitment moves third-party frontier-model evaluation from periodic access to an intended standing oversight function that can inspect training processes and incidents, not only completed models. If implemented as described, it would create a more…