Google's AI Solves 10-Year Problem in Just Two Days

In essence, the system was able to uncover a previously unknown mechanism of horizontal gene transfer.

The Google AI tool was able to reveal the mechanism of antibiotic resistance transmission in superbugs in just two days, a problem that has taken researchers at Imperial College London 10 years to study.

Scientists led by microbiology professor José Penades studied cf-PICIs bacteriophages, which can infect various types of bacteria. They hypothesized that some viruses borrow "tails" from others to insert their DNA into target bacterial cells. Experiments confirmed the hypothesis, revealing a previously unknown mechanism of horizontal gene transfer.

The researchers then decided to test the Google AI system designed to assist in scientific research. They formulated a question, and within two days, the algorithm provided answer options—one of which was their own hypothesis. This indicated that the AI independently analyzed the available data, excluded "dead-end" theories, and derived the key idea without experiments.

Although this tool does not replace laboratory research, it can significantly accelerate the scientific process by eliminating unnecessary hypotheses. However, the application of AI in science remains controversial: some works created with its involvement turn out to be irreproducible or unreliable. Therefore, scientists are working on methods to verify the accuracy of conclusions and establish ethical standards for AI research.