Who can apply for OpenAI's free ChatGPT grants?
OpenAI will hand out 50,000 more free ChatGPT grants to academic researchers, the company announced Monday, as it pushes to reach 200,000 researchers in its program. Applications open November 23. Unlike the first round, the grants will go to the first eligible applicants instead of being drawn by lottery.
What changed between OpenAI's pilot and the new ChatGPT grants?
The catch is the subscription tier. Researchers in the pilot received ChatGPT Pro, OpenAI's top plan. The new grants come with ChatGPT Plus, the cheaper option. OpenAI says most pilot users would have stayed within Plus limits anyway, and the switch lets the company reach more teams.
Each new grant covers a researcher and up to two collaborators, or up to 150,000 people in total across the 50,000 grants. That is fewer collaborators than the pilot allowed. In the original program, each researcher could invite up to four collaborators.
The pilot launched in July, and OpenAI now aims to double the number of researchers it set out to reach at that time. So far, 20,000 researchers in 33 countries have joined.
OpenAI also published a study of how more than 1,800 scientists used ChatGPT during the pilot. On average, reading and learning made up 28% of their science messages, and writing made up 27%. Survey participants said the same work would have taken nearly twice as long without AI.
Does OpenAI's study prove AI speeds up scientific discovery?
That figure is self-reported, and OpenAI says the study does not prove AI speeds up discovery. The company also named funding, staff and lab capacity as limits on how quickly researchers can test new ideas. In other words, faster writing and reading does not automatically translate into faster experiments.
The grants are one piece of a broader push into science. OpenAI offers free seats on GPT-Rosalind, its biology model, to US university labs. It also gave $5 million to a Simons Foundation effort to help mathematicians check AI results. In August, the company said its unreleased Astra model had claimed to solve ten open problems in math.
The competition for scientific users is heating up. Anthropic launched Claude Science, an AI workbench designed for labs, in June. In October, Google DeepMind opened its science models to 200 European researchers. For universities and research groups deciding which tools to adopt, the offers are arriving fast, and several of them are free for the first wave of users.
Dieser Artikel wurde mit Unterstützung von KI verfasst.
News Factory APP - agentische News für besseres SEO & AEO.
