Google’s AI solved a problem that scientists have been struggling with for more than 10 years in 2 days
In two days, Google’s AI tool was able to solve a complex scientific problem related to the resistance of some superbugs to antibiotics, which takes millions of lives every year. Researchers from Imperial College London spent more than ten years solving this problem, informs Live Science.
A team of scientists led by Professor of Microbiology Jose Penades studied bacteriophages cf-PICIs capable of infecting various types of bacteria. They hypothesized that some viruses can use the “tails” of others to introduce their DNA into host bacterial cells. Experimental data confirmed this hypothesis, revealing a new mechanism of horizontal gene transfer that was previously unknown to the scientific community.
Before publishing their results, the scientists turned to the AI tool “Google AI co-scientist”, specially designed to support scientific research. In two days, the algorithm offered several answer options, one of which completely coincided with the researchers’ hypothesis.
“This effectively meant that the algorithm independently analyzed the available data, rejected the false theories and proposed the same hypothesis that we arrived at after years of painstaking work, but much faster.” Professor Penades noted.
The scientists emphasized that the use of AI does not eliminate the need for experimental verification, but it allows to speed up the formation of hypotheses, saving years of research.
