AI's role in science sparks debate as firms race to harness biological data
As AI firms collect vast biological datasets and claim breakthroughs, mathematicians and journalists raise concerns about transparency and scientific integrity.
At a glance
- Basecamp Research stores 100 billion genes in London to train AI systems for biological research.
- OpenAI claimed its AI solved a Navier-Stokes problem, raising concerns from mathematicians about research overlap.
- Google DeepMind's AlphaFold predicted 200 million protein structures, earning a Nobel Prize in Chemistry.
A facility in central London stores 100 billion genes in freezers, part of an effort by Basecamp Research to gather biological data for AI systems. The company aims to build artificial intelligence that understands biology, mirroring how tech firms have used online text to create large language models.
Dr Oliver Vince, co-founder of Basecamp Research, believes AI will accelerate scientific discovery. However, recent events highlight growing tensions in the field.
Last month, OpenAI announced its AI had solved a longstanding mathematics problem related to the Navier-Stokes equations. The announcement came just hours after mathematicians Tristan Buckmaster of New York University and Levent Alpöge of Anthropic raised concerns about potential conflicts. Buckmaster questioned whether their joint research, stored in OpenAI's Codex tool, may have influenced OpenAI's work, though he stopped short of making direct accusations.
Scientific publishers are also grappling with the rise of AI-generated content. Dalmeet Singh Chawla, a science journalist, warns that generative AI is disrupting the peer review process. He argues that publishers' integrity teams are understaffed to handle the volume of AI-assisted submissions.
Despite these concerns, AI has shown remarkable promise in scientific research. Google DeepMind's AlphaFold, released in 2020, predicted 200 million protein structures in a year—work that would have taken millions of years experimentally. The breakthrough earned DeepMind leaders a Nobel Prize in Chemistry.
Dr Pushmeet Kohli of DeepMind described AlphaFold as a watershed moment for biology, demonstrating AI's potential to transform scientific research.
Source: BBC Science
Questions and answers
- What is Basecamp Research doing with genetic data?
- Basecamp Research collects DNA samples stored in freezers in London, aiming to gather enough biological data to create AI that can understand the language of biology, similar to how large language models are trained on text.
- What controversy surrounds OpenAI's math breakthrough claim?
- Mathematicians Tristan Buckmaster and Levent Alpöge raised concerns that their joint research stored in OpenAI's Codex tool may have influenced OpenAI's claimed solution to a Navier-Stokes problem, though no direct accusations were made.
- How has AI already impacted biological research?
- Google DeepMind's AlphaFold predicted 200 million protein structures in one year, work that would have taken millions of years experimentally, earning its developers a Nobel Prize in Chemistry.
