AIが期間内の動向を整理
Anthropic and OpenAI updates move AI forward on both use cases and manageability
In the AI news confirmed on Sept. 20, Anthropic expanded life sciences use cases and research support, while OpenAI announced updates that make ChatGPT easier to manage and use for ads. For business users, the focus is increasingly not only on expanding AI usage, but also on measuring value and handling it safely.
SOURCE CHECK
Primary Sources 8
Primary Sources
Key Points
- 1Anthropic expanded Claude usability for life sciences
- 2Claude Science improved biomolecular model speed and memory use
- 3OpenAI made usage, cost, and outcome tracking easier in ChatGPT Admin Console
- 4ChatGPT Ads became easier to start through external tool integrations
- 5There were also updates around safety and evaluation tracking
Anthropic expanded life sciences use
Anthropic launched the Life Sciences Verification Program, creating an environment where life sciences professionals can use Claude more easily. The key point is that it broadens use across research, drug discovery, clinical development, and manufacturing. It is an update that helps AI move from experimentation into day-to-day work.
Claude Science lowers research workflow friction
Claude Science supported optimization of more than 30 open-source models, with an average of about 4x speed improvement and lower memory use. For biomolecular modeling, heavy processing is often a barrier to adoption, so these operational improvements matter in practice.
OpenAI improved visibility into value
ChatGPT Admin Console now makes it easier to review usage, spend, tasks, and outcomes together. Administrators can better understand which teams are using AI and for what, making it easier to explain value and decide how to roll out adoption further.
Ad workflows and safety checks also moved forward
OpenAI also outlined new ways to buy ChatGPT Ads and integrations with Shopify and HubSpot, aiming to make campaigns easier to start and measure. In parallel, it published a framework for reporting model misalignment, continuing the push to identify safety issues earlier.
Related News Items
Anthropic / availability
Claude becomes easier to use for life sciences work
Anthropic / feature
Run biomolecular modeling faster and with less memory
OpenAI / feature
See AI spend and outcomes in one place
OpenAI / feature
Advertisers can launch campaigns with fewer steps
Anthropic / availability
Track Anthropic’s latest updates and safety work in one place
OpenAI / policy
Share model misalignment early to support safety checks