Google DeepMind CEO: AI-Designed Drugs Coming to Clinical Trials in 2025

Nobel laureate and Google DeepMind CEO Demis Hassabis said Tuesday (Jan. 21) that he expects to see pharmaceutical drugs designed by artificial intelligence (AI) to be in clinical trials by the end of the year.

During a fireside chat at the World Economic Forum in Davos, Switzerland, Hassabis said these drugs are being developed at Isomorphic Labs, a for-profit venture created by Google parent firm Alphabet in 2021 that was tasked to reinvent the entire drug discovery process based on first principles and led by AI.

“That’s the plan,” Hassabis said.

While large language models have taken the spotlight, Hassabis said that the field of AI as it applies to science is “a lot richer than just the language models and things like AlphaFold.” AlphaFold2, for which he and colleague John Jumper won the Nobel Prize in Chemistry in 2024, is an AI model that predicts the 3D structure of proteins, solving a half-century biology challenge.

Ardem Patapoutian, a professor of neuroscience at Scripps Research and a 2021 Nobel laureate in Physiology or Medicine who joined Hassabis at the panel, described AlphaFold as “one of the most amazing, quick advancements in science I’ve ever experienced.” He said 25 years ago, it took a Ph.D. student five years to find out the structure of a protein. With AlphaFold, “just type in the sequence and it tells you the structure,” he said.

Hassabis said AlphaFold has now predicted the structures of 200 million proteins known to science — work that would have taken an estimated billion years using traditional methods. The latest version, AlphaFold3, has expanded capabilities to analyze protein interactions with other proteins, ligands and DNA/RNA.

AI Comes Full Circle

Patapoutian also highlighted AI’s potential to unlock the mysteries of brain function, particularly in understanding complex neural patterns and their relationship to behavior. He noted that while current technology can predict behavior in simple organisms, understanding more complex brains remains a significant challenge that AI could help solve.

“Overall, neuroscience is very excited about AI,” Patapoutian said. It could help scientists make headway in understanding the brain because “despite decades of research, we still really don’t understand how the brain works.”

Even after looking at the pattern of neurons firing in a brain, it remains difficult to predict what behavior is going to come next. Perhaps for a simple creature like the C. elegans worm that only has 300 neurons, it is possible, Patapoutian said. But for anything more complex, “we have absolutely no idea, and that has been one of the Holy Grails of neuroscience, not just to predict behavior, but (also) more complex thoughts, intelligence, consciousness.”

Hassabis said in this sense AI has come full circle. The structure of the brain inspired AI’s neural networks, and now AI can help scientists understand how the brain works.

Hassabis said the next frontier for AlphaFold3 is to determine how mutations can cause changes in the structure and function of the protein. AlphaFold2 solved the problem of a static protein, but he pointed out that proteins are not static. Eventually, he sees the advent of personalized medicine, where a drug is optimized for each individual’s own metabolism.

Virtual Cells and AGI

Looking ahead, Hassabis outlined his vision for a “virtual cell” simulation that could revolutionize biological research. Patapoutian said traditionally, the way one finds a protein structure is to pull it out of a cell. But then you don’t know its natural position in the cell. Seeing the cell as a whole and where proteins are located would be more informative, Patapoutian added.

For example, being able to view a protein in a cell could let you see that it is, say, localized at the tip of the neuron where specific activity is going on, Patapoutian said. “That’s going to give you a very different understanding than just levels of expression, for example.”

Patapoutian also wondered where Hassabis procured the data to train his AI models. Hassabis said beyond the public datasets, there are companies you can hire to generate specialized data to fill in the gaps. His team also uses synthetic data they generate themselves. On the flip side, they could also develop algorithms that need less data to train on. This is to mimic human beings, who can generalize based on a few examples.

Asked when artificial general intelligence (AGI) will arrive, Hassabis said people who say it will arrive in months might have an ulterior motive of needing to raise funding. “AI is over-hyped in the near term, but I think it’s still under-appreciated in the medium- to long-term.” He believes AGI could come in five to 10 years, but a couple of major breakthroughs need to arrive first.

These breakthroughs are a fully reasoning and planning skillset, and the ability to truly be creative, not just mimic artistic styles or novel ideas it has been trained on, Hassabis said. “Could you come up with general relativity like Einstein did based on the knowledge that he had at the time in the 1900s? I don’t think any of our systems could do anywhere close to that.”

The post Google DeepMind CEO: AI-Designed Drugs Coming to Clinical Trials in 2025 appeared first on PYMNTS.com.