A tough problem stumped microbiologists for ten years. They finally solved it. Then, a new artificial intelligence (AI) tool figured it out in just two days, BBC News reports. José R. Penadés and his team at Imperial College London studied why some superbugs resist antibiotics.
Penadés tested a tool called “co-scientist,” made by Google. He gave it a short question about his research. In 48 hours, it matched his team’s conclusion. His work wasn’t public, so he was stunned. He told the BBC he needed an hour alone to process it. He even emailed Google, asking if they accessed his computer. They said no.
His team’s decade of work included proving their idea. That alone took years. With the AI’s help from the start, they could’ve saved time.
AI is a powerful helper
We "challenged AI co-scientist to generate scientific hypotheses," Penadés posted to X. "It correctly predicted how cf-PICIs spread between bacterial species, matching our unpublished, experimentally confirmed mechanism!" A preprint published on bioRxiv tells the full story.
Professor Penadés said the tool did more than copy his findings. It offered four other ideas. All made sense. One was brand new, and his team is now exploring it. The superbug problem they tackled involves how these germs spread. Their idea, or hypothesis, is a possible explanation. They think superbugs use virus tails to move between species. He compared it to having “keys” to unlock new hosts.
This idea was secret, unique to his team. No one had shared it. So, he tested Google’s AI with it. Two days later, it nailed the tail idea as its top answer. The team was amazed.
People debate AI’s impact. Some say it boosts science. Others fear it cuts jobs. Penadés gets the job worries. But he sees AI as a powerful helper. His team believes it’ll shape the future. He feels thrilled to use it. To him, it’s like playing a huge game and winning. This tool could change how science works, he says. It’s fast, smart, and opens new doors.