OpenAI
-
Research
OpenAI unveils LifeVal, an internal benchmark that grades AI agents on everyday errands such as childcare, budgets and government paperwork
OpenAI has published LifeVal, an evaluation of practical personal tasks drawn from time-use studies and a survey of 3,995 people, which it says has guided a year of work on browser use, computer use and connectors. CNN r…
-
Infrastructure & Chips
OpenAI launches Grid Protectors Program, pledging that its future US compute campuses will cut power use when grids are strained
OpenAI has launched a national Grid Protectors Program under which its future compute campuses will agree with utilities how much electricity demand they can shed, how quickly and for how long. The pilot is its planned E…
-
New Products
OpenAI will merge ChatGPT's Chat and Work tabs and pool ChatGPT and Codex usage into one allowance from December
OpenAI told ChatGPT subscribers it will test a combined Chat and Work experience with a small group of users before a wider rollout, while Codex stays separate. Its Help Center says that from December, Plus, Pro and Busi…
-
Business
OpenAI trims its funding round target to about $20 billion after a cool reception from investors, reports say
OpenAI is now aiming for roughly $20 billion in its pre-IPO round instead of the $30 billion it was chasing a week ago, according to the Financial Times and CNBC, while the Wall Street Journal puts commitments so far nea…
-
Policy & Regulation
Amicus briefs pile up in the OpenAI copyright fight as a federal judge prepares to weigh fair use
With summary judgment motions pending in the consolidated copyright cases against OpenAI and Microsoft, OpenAI on October 30 published excerpts from friend-of-the-court briefs that back its fair-use defense. Publishers a…
-
Policy & Regulation
OpenAI, Google, Microsoft and Anthropic back one robots.txt signal for opting out of AI training
OpenAI and Google said on October 30 that they will honor a shared robots.txt token, drafted in the IETF's AI Preferences effort, that lets a website state its AI-training choice once for several providers. Microsoft and…
-
Safety & Ethics
OpenAI tightens ChatGPT's violence safeguards, limits weapons information for teens and promises a twice-yearly safety report
OpenAI said on October 30 that ChatGPT will give less detail about past mass attacks, may end chats that dwell on harming others and will further restrict weapons-related answers for users it believes are under 18. The c…
-
Business
Thrive Holdings, the OpenAI-backed buyer of traditional companies, agrees to take ADT private in a $15.3 billion deal
ADT agreed on October 29 to be acquired by Thrive Holdings for $10 a share in cash, about $7.6 billion in equity and $15.3 billion including debt. The deal, financed by Apollo-managed funds, is the largest yet for the Th…
-
AI Governance
OpenAI rewrites its Model Spec around five values, putting faithfulness to human intent at the core of how GPT should behave
OpenAI published a revised Model Spec on October 29 that organizes its rules for model behavior around five values, with faithfulness to human intent first, followed by avoiding harm, honesty and transparency, support fo…
-
Safety & Ethics
OpenAI widens its misalignment monitoring to cover frontier model training, evaluations and some customer deployments
OpenAI said on October 29 that it has substantially expanded the system it uses to catch its most capable models misbehaving, after the Hugging Face breach began in evaluations it was not watching. The four-stage pipelin…