AIが期間内の動向を整理
OpenAI and Anthropic updates: legal AI, government discounts, and transparency metrics are released
On Sept. 18, OpenAI unveiled Astra for Law and government adoption support, while Anthropic published transparency metrics and misuse cases. For companies using AI in practice, these updates matter not only for new capabilities but also for adoption terms and safety checks.
SOURCE CHECK
Primary Sources 12
Primary Sources
Key Points
- 1OpenAI announced Astra for Law, with a Legal Search Index covering over 230 million URLs and 26 plugins for legal workflows.
- 2OpenAI also announced discounted access and adoption support for U.S. government agencies.
- 3Anthropic published three metrics on AI R&D share, agent oversight, and compute allocation.
- 4Anthropic also shared misuse cases and information on coding- and knowledge-work-focused models.
What happened
OpenAI announced Astra for Law and showed a system for searching a broad set of legal materials, plus partner plugins and workflow settings for law firms. It also announced discounted access and adoption support for government agencies.
Why it matters
In legal research and public-sector adoption, usability, governance, and cost matter as much as raw capability. This update shows that AI is moving deeper into concrete business workflows.
What Anthropic disclosed
Anthropic published three metrics on the share of R&D done by AI, oversight of AI agents, and compute allocation, and it also shared misuse cases. Transparency and safety are becoming more visible comparison points.
What this means for work
For enterprise decisions, it is important to evaluate not only features, but also how confidential data is handled, what oversight is in place, and how much adoption will cost. These announcements help frame those decision criteria.