Vivodyne Develops Robotic Labs to Generate Human Tissue Data for AI Drug Discovery

Addressing the AI Drug Discovery Data Gap

Vivodyne, a biotech startup, claims the AI drug-discovery industry faces a significant data problem. The company has developed a solution called HIVE, which consists of modular robotic labs designed to generate relevant biological data.

HIVE can grow 20 types of human tissue and autonomously dose and monitor them. This process creates causal biological data that is currently missing from AI models, which often rely on data from animal testing, single cells, or proteins.

Critique of Current AI Drug Discovery Claims

Andrei Georgescu, Vivodyne’s CEO and co-founder, highlighted that without human testing data, AI models are more likely to find cures for diseases in mice rather than humans. This perspective aligns with recent comments from Anthropic CEO Dario Amodei, who noted that claims of AI curing cancer have become more cliché than credible.

Despite ambitious statements from figures like Sam Altman and Demis Hassabis regarding AI's potential to cure diseases, actual results in AI-designed drugs have been limited. Only a few AI-designed drugs have reached human trials, with one progressing to Phase III.

Limitations of Existing AI Models and Data

Nobel-prize winning Alphafold, while a significant advance in understanding life's building blocks, has not yet produced a new drug. Isomorphic Labs, founded to build on Alphafold, anticipates its first trials by the end of this year, emphasizing the need for highly accurate predictive models across a wide range of biochemical properties.

Georgescu argues that existing AI models lack the data necessary to capture the complexity of human biology. This challenge is also evident in the pharmaceutical industry, where 90% of drugs effective in animal testing fail to receive regulatory approval for human use.

Vivodyne's Approach to Human Tissue Modeling

Vivodyne, spun out of the University of Pennsylvania in 2021, aims to provide a different approach. The company states that its tissues closely mimic the behavior of real human organs. For example, its liver cells demonstrate 94% predictive accuracy when compared to human trials for toxicity testing.