by 
Google has unveiled a new system called “Co-scientists”, from which indicates that it could lead to big new breakthroughs.
The artificially intelligent tool is operated by similar technologies as in chat-based major language models such as the Gemini own Gemini and competitors such as chatt. However, it is specifically dedicated to finding new research documents.
Scientists who used it describe it as a particularly well -read and helpful colleague. It has already been used by researchers at Imperial College London, who were able to reproduce their own work much faster than they would have done without the system.
To use the system, an experienced scientist uses a normal language to give it a research goal. The system then checks the published literature and synthesized it, but can also evaluate its own results and suggest new hypotheses and possible experiments that could be validated – all while quoting literature and explaining its suggestions.
According to Google, the system is specially created as collaborative instead of automating research completely via AI. Scientists can talk to him, give feedback and change their understanding that it could improve over time.
The company has carried out a number of options for evaluating the system, including the request of experts to achieve 15 “challenging and open research goals in their field” and to produce the co-scientist possible solutions. These experts rated the new tool about other existing tools.
Scientists speculated that the system could be a new way of quickly finding information about new hypotheses and new breakthroughs on a number of important topics. It has already been used to deal with research on antimicrobial resistance, for example what the World Health Organization says that it is one of the greatest threads for global well -being and security.
“Laboratory science is resource -intensive. With global challenges such as antimicrobial resistance, it is clear that we do more with less and accelerate new discoveries”.
“When the Google Research Team turned to us to test its KI platform, we found that we had to edit it with the same scientific questions that we had already examined ourselves and used it as the basis for our experimental work .
“This effectively meant that the algorithm was able to examine the available evidence, analyze the possibilities, ask questions, to design experiments and to suggest the same hypothesis.
