Unlock free Digest editor
Roula Khalaf, editor of FT, chooses her favorite story in this weekly newsletter.
Google has built a laboratory assistant for artificial intelligence to help scientists accelerate biomedical research, while companies are racing to create specialized applications from top technology.
The so-called co-scientific tool of the US technological group helps researchers to recognize flaws in their knowledge and propose new ideas that could accelerate scientific discovery.
“What we are trying to do with our project is to see if technology like AI co-scientificists can provide these researchers with superpowers,” said Alan Karthikesalingam, a higher staff scientist on Google.
GoogleThe new tool comes because technological companies spend billions of dollars on AI models and products, believing that technology can change industries from health care in energy and education.
Openii, Confusion and German BIONTECH MEDICINE and its AI branch based in London have recently launched their own possess a AI research toolsWhile Alphafold Google Deepmind has shown that fast developing technology can accelerate scientific research.
Early tests of Google’s new tool with experts from Stanford University, Imperial College London and Houston Methodist Hospital revealed that it is able to create scientific hypotheses that showed promising results.
The tool managed to make the same conclusions – for a new mechanism for the transmission of genes that helps scientists understand the spread of antimicrobial resistance – as a new breakthrough of researchers to Imperial.
The results of the imperials were not in a public domain because they were reviewed in a top scientific journal. This has shown that Google’s co-scientific tool has been able to achieve the same hypothesis using Ai resonating in just a few days, compared to the age that the University team has spent exploring the problem.
The AI tool could also help researchers at Stanford to find existing drugs that could be remodeled to treat liver fibrosis, a serious illness in which a scar is accumulated in the organ. Google’s co-scientist suggested two types of drugs for which Stanford scientists have found to have helped treat the disease.
“We think it will be a tool that has the potential to change the way we are approaching science,” said José Panadés, a professor at the Imperial Department of Infectious Diseases and Fleming Initiatives, which was part of the team behind the study of the new Transmission Mechanism gene.
The tool operates using several AI agents that mimic the scientific process. For example, one AI agent specializes in generating ideas and the other in thinking and reviewing these ideas, said Vivek Natarajan, a research scientist at Google.
The model is able to download information from scientific papers and special databases that are freely available online and other tools, such as Alphafold. He then analyzes the information given to him and presents researchers with a ranked list of proposals with explanations and connections to sources. Researchers can then purify these proposals.
Tools such as Google’s AI co-scientist could help scientists be up to date with all the new information generated in their fields, said Jakob Foerster, an associate professor at the University of Oxford, who also developed AI tools for research. “I think it’s super valuable,” he said.
