Follow ZDNET: Add us as a favorite source On Google.
Key takeaways of zdnet
- Openai for science was announced in Tuesday x post.
- Its goal is to accelerate scientific discovery through AI.
- The post suggests that the GPT-5 will play an important role in the attempt.
Artificial intelligence researchers have long dreamed of automatic scientific discovery process. Now Openai is ready to turn that vision into reality.
According to Tuesday, the company is taking an initiative called OpenIAI for Science for the purpose of manufacturing “The Next Great Scientific Instrument: AA-AI-Powered Platform, which accelerates scientific discovery. X post From Kevin Veal, the company’s chief product officer. The timeline for the project is not yet known; Wayel said in his post that more information will come in the coming months.
Also: Each AI model is removing the drug – and Lamrena proposes a fix.
(Disclosure: ZDNET’s original company Ziff Davis filed a case of April 2025 against Openai, alleging that it violates Ziff Davis copyright training and operating its AI system.)
Wayle, who would lead the effort, wrote that the openiE will appoint a team of “world class” academics that are “fully AI-filled” and “great science communicators” to join a small group of already employed researchers already employed by the company.
What can we expect from the project?
There is not much to be known about the “platform” at this point that the Vail and their runaway team plan to plan. But their X post indicates on some aspects of the scientific process that OpenIAI for science may try to automatically automatically automatically.
Wayel wrote that the newest model of OpenaiI, GPT -5, which was released last month, is “clearly a new border for AI” to help in advance scientific progress for AI. For example, he recently cited theoretical physics paper In which the model was used to suggest ideas for evidence, suggesting that openi for science, eventually, in part, can help researchers to formulate and prepare research methods, which speeds up the speed of search.
Also: Doctors trust AI medical advice on patients – even when it is wrong, it is still found.
Wayel’s X Post Smack Interesting GPT -5 Smack of a sales strategy: The model has received mixed reviews as it was launched, many users complained that it was inferior to its predecessor, GPT -4O. With the company’s new scientific research program, by wedding GPT-5, the Openai can try to fix the damaged reliability of the model. Ultimately, if the company can prove that the GPT-5 is capable of contributing meaningfully to scientific discovery-a rigorous process that requires abstract, multistap logic, and which can result in practically practical advantage for society on a large scale-may be more interesting to rely on models with their own sensitive works.
Openai did not immediately respond to ZDNET’s request for comments.
While Wayle did not mention grant-lighting in her blog post, this is another area in which generic AI tools like chatgip can be applied fruitfully: researchers currently spend close to half of half (45%) of their time Writing grant proposals, institute for progress according to the think tank.
Great progress
AI has so far discovered new physical laws, treated cancer, suggested a comprehensive solution for climate change, or made any of other major scientific successes that many artificial general intelligence can be around the corner. Perhaps one day an AI model will be able to fully automate the scientific process, fully preparing new hypotheses to conducting experiments to analyze results. But for some time, AI’s scientific skill lies in the ability to identify the complex network of patterns from existing data.
Nevertheless, researchers have made some significant progress, and AI is rapidly becoming an integral tool in mainstream science.
Also: Open AI, invite anthropic American scientists to experiment with frontier models.
Google Deepmind CEO Demis Hasabis and director John Jumper honored Nobel Prize in Chemistry Last year for their work on Alphafold2, which uses AI to predict the structure of almost all known proteins. Nobel Prize in Physics Last year, Geoffrey Hinton was awarded, provided for the so -called “AI’s Godfather” and his leading work on the nerve network, one of the physicist John Hopfield, which became a technical outline for the current AI boom.
AI’s mathematical capabilities are also developing rapidly. In July, Openai reported that an experimental logic model of it achieved a gold medal level performance on the International Mathematics Olympiad, which is widely considered one of the world’s most elite mathematics competitions. Google Deepmind reported the same level of performance from its own model, Gemini 2.5 Pro.
Want to follow my work? Add ZDNET as a reliable source on google,
