According to a Bloomberg report on Tuesday (21st), Isomorphic Labs, an AI-driven drug discovery startup under Google DeepMind, plans to initiate clinical trials for its AI-designed drugs by the end of 2025. The company's CEO, Nobel Chemistry Prize laureate Demis Hassabis, stated during a World Economic Forum (WEF) panel discussion, "We aim to start clinical trials for AI-designed drugs by the end of this year. That is our goal."
Isomorphic Labs is committed to drastically reducing the time required for drug discovery, compressing what traditionally takes over a decade into mere weeks or months. Hassabis, alongside fellow DeepMind scientist John Jumper and a U.S.-based professor, won the 2024 Nobel Prize in Chemistry for their groundbreaking research in protein structure prediction.
AI’s ability to process massive datasets offers significant potential to accelerate drug development. However, according to Bloomberg Intelligence analyst Andrew Galler in a December 2023 report, major pharmaceutical companies remain cautious about adopting this emerging technology, citing mixed results from early-stage clinical candidate data.
Despite this, collaborations between tech companies and pharmaceutical giants are increasing. Last year, Isomorphic Labs entered strategic research agreements with two leading pharmaceutical companies, Eli Lilly and Novartis.
Founded by Alphabet in 2021, Isomorphic Labs was established to commercialize DeepMind's AI technologies, focusing on drug discovery. In 2018, DeepMind launched the AlphaFold tool to predict protein structures. Now in its third generation, AlphaFold can simulate various molecular structures, including DNA and RNA, and predict their interactions.
The application of AI in drug design is accelerating the drug development process, despite ongoing challenges. The potential for future advancements remains significant.
GPT-4b Micro: Revolutionizing Protein Programming and Stem Cell Research
Through protein reprogramming techniques, GPT-4b Micro has enhanced stem cell generation efficiency by 50-fold. The AI-optimized proteins exhibit far superior capabilities in transforming ordinary cells into stem cells compared to their natural counterparts.
Unlike Google DeepMind's AlphaFold, GPT-4b Micro acts as a "protein interaction expert," accurately predicting interactions between different proteins. These proteins play critical roles in fetal development, effectively serving as a "time machine" for life—rewinding mature cells to a youthful state, offering limitless potential for stem cells.
This groundbreaking AI model opens new possibilities for treating diseases once considered incurable, such as diabetes and heart disease.
Practically, AI-driven approaches may bring transformative breakthroughs in organ transplantation. Scientists could potentially tailor organs for individual patients, significantly reducing the risk of rejection and improving transplant success rates.
GPT-4b Micro also marks OpenAI's first formal announcement of its AI model achieving revolutionary scientific discoveries. This reinforces OpenAI CEO Sam Altman’s claim that he is "confident" the company has found a pathway to constructing AGI (Artificial General Intelligence). Altman believes superintelligent tools could dramatically accelerate scientific discovery and innovation, far surpassing human capabilities.
From AI to nuclear fusion energy and life extension, Altman is optimistic about addressing humanity’s most pressing challenges.
Notably, the new model is the result of a collaboration between OpenAI and Retro Biosciences, a company in which Altman invested $180 million. Currently, the model is not publicly available, and no timeline has been set for its commercialization.
Resource: DeepMind預估今年啟動AI設計藥物臨床試驗
