Isomorphic Labs’ AI Ready for First Human Trial of an Anti-Cancer Drug | Microsoft ai copilot | Mckinsey generative ai report pdf | Generative ai benefits for business examples | Turtles AI

Isomorphic Labs (Alphabet/DeepMind) is preparing to launch its first fully AI-designed cancer drug in 2025, leveraging AlphaFold3 to predict complex protein structures. Strong partnerships (Novartis, EliLilly) and $600 million in investments are accelerating its development. Amid expectations and concerns about transparency, accessibility, and accountability, the upcoming clinical trial represents a breakthrough in drug design.

Key Points:

• Imminent launch of the first human clinical trials of an AI-designed cancer drug.
• Use of AlphaFold3 to model protein structures and complex molecular interactions.
• Collaborations with Novartis, Eli Lilly, and $600 million in funding from Thrive Capital.
• Open questions about patents, costs, legal liability, and transparency of the AI ​​"black box."

Alphabet, through its subsidiary IsomorphicLabs (formed from DeepMind in 2021), is preparing to test an entirely AI-powered anti-cancer drug for the first time in humans in the coming months. This was confirmed by Colin Murdoch, president of Isomorphic Labs, in an interview with Fortune, and reiterated in official communications: "We are very close to administering the drug to the first patients." The project uses AlphaFold3, the latest AI system capable of predicting three-dimensional protein structures and interactions with molecules such as DNA, RNA, and ligands.

The company, headquartered in London with additional offices in Switzerland, has signed multi-billion dollar deals with Novartis and Eli Lilly, and raised $600 million in a round led by Thrive Capital. These funds will be used to develop the AI ​​drug design engine and fund internal oncology and immunology programs. The trials will begin soon, supported by a strengthening of the internal clinical management team.

AlphaFold3, released in May 2024 and now available for academic use, represents a significant advancement over the previous model: it goes beyond predicting globular structures to model multiprotein complexes, ligand binding, and post-translational modifications, demonstrating improvements of up to 50% compared to traditional technologies. This precision accelerates rational compound design, reducing time and costs compared to the typical 10–15 years and over a billion dollars required for a traditional drug.

However, these ambitions raise significant questions: who holds the patents on AI drugs? Will they be accessible and not prohibitively expensive? What legal liabilities arise if a drug generated by a "black box" fails? And will regulatory rigor keep pace with the speed of in silico discoveries?

Throughout all of this, Isomorphic Labs remains private: "We have nothing further to add," a spokesperson told Gizmodo. The scientific community and the public await answers on transparency and governance as the first human doses designed by algorithms approach. This unprecedented and delicate phase could open new avenues in medicine—or highlight limitations to be overcome.

The start of human trials represents a critical step: AI could offer decisive advantages, but it raises unanswered questions about the role, responsibility, and social impact of drugs designed by non-human systems.