news
Anthropic's AI Enzyme Discovery Under Scrutiny for Prior Research
AI giant Anthropic is claiming a novel enzyme discovery using its AI agents in a new biolaboratory. However, some researchers, including former University of Copenhagen PhD student Mario Rodríguez Mestre, are questioning the originality, citing previous similar findings. This raises questions about what constitutes an 'AI discovery' and the ethics of attributing scientific breakthroughs.
AI giant Anthropic is claiming a novel enzyme discovery using its AI agents in a new biolaboratory. However, some researchers, including former University of Copenhagen PhD student Mario Rodríguez Mestre, are questioning the originality, citing previous similar findings. This raises questions about what constitutes an 'AI discovery' and the ethics of attributing scientific breakthroughs.
The architecture take
Anthropic's shiny new AI 'discovery' of enzymes is facing the oldest problem in science: prior art. When AI 'finds' something a human already published, it forces us to ask if the AI is truly discovering or just, well, reading. If the AI was trained on data that included Mestre’s earlier work, then this isn't a discovery. It's a very expensive re-discovery, or worse, an algorithmic echo of someone else's labor.
This isn't just about credit; it's about the nervous system load of actual scientists who spend years at the bench. Imagine dedicating your career to a mechanism, only for an algorithm to 'discover' it after ingesting your data. The real mechanism here is recognition and the dopamine hit that comes from a novel insight. If AI takes that, it changes the entire reward system of research. What happens to human innovation when the digital archive becomes the 'discoverer'? I'm watching to see how this plays out for patents and research funding — because the precedent set here matters for every lab working on enzymes, peptides, and new biologicals.