OpenAI announced on Monday an extension of its system intended for academic research. The company intends to reach more teams thanks to a less expensive formula, without presenting the declared time savings as proof of acceleration of discoveries.
The scientists interviewed declare time saved
According to an OpenAI study of more than 1,800 scientists who participated in the pilot program, reading and learning activities accounted for an average of 28% of science-related messages, while writing accounted for 27%. Survey participants estimate that the same work would have taken almost twice as long without AI, but this estimate is based on their own statements. OpenAI specifies that this study does not demonstrate that AI accelerates scientific discovery and mentions that funding, staffing and laboratory capacity remain limits to the experimentation of new ideas.
OpenAI plans 50,000 helpers with access to Plus
OpenAI intends to distribute 50,000 new free grants to academic researchers, with each grant covering one researcher and up to two collaborators, representing a total of 150,000 people. Registrations will be open from November 23 and aid will be awarded to the first candidates meeting the eligibility criteria, without resorting to a drawing system. Those selected will receive access to ChatGPT Plus, while the pilot program offered ChatGPT Pro, a more expensive version. According to OpenAI, the majority of pilot participants would have used less than Plus’s capabilities, which helps increase the number of teams that can benefit from it. To date, the program brings together 20,000 researchers spread across 33 countries. OpenAI now aims to reach 200,000 researchers, twice the initial objective set when the pilot was launched in July, during which each researcher had the possibility of inviting up to four collaborators.
The evening letterFree · 7 p.m.
Are you following the race for AI models?
Each output (GPT, Claude, Gemini, Mistral, etc.) decrypted the same evening, in 5 minutes. Free.
Included upon registration: our selection of the best AI guides & comparisons.
Choose your pace
Free · No spam · Unsubscribe in 1 click
Anthropic also offers a tool to laboratories
Other programs are also aimed at the scientific community. In June, Anthropic presented Claude Science, an AI experimentation platform intended for laboratories. In October, Google DeepMind made its scientific models available to 200 researchers in Europe. OpenAI also offers free access to GPT-Rosalind, its biology model, for US university laboratories, and has awarded $5 million to a Simons Foundation initiative to help mathematicians validate results generated by AI. In August, Astra, an OpenAI model that has not been released, announced that it had solved ten unsolved mathematical problems.

