The Einstein test: what happens when AI tries to rediscover relativity?
Researchers and AI leaders are debating whether large language models can achieve artificial general intelligence by attempting to rediscover complex scientific theories like Einstein's general relativity. Demis Hassabis of Google DeepMind suggests using historical data as a benchmark for AI capability.
Why it matters
This experiment serves as a high-level test to determine if AI can perform genuine scientific discovery rather than just pattern matching.
Philip Ball is a science writer based in London.
Search author on: PubMed Google Scholar
Bluesky Facebook LinkedIn Reddit Whatsapp X Save article View saved research Illustration: Acapulco Studio
In 1915, Albert Einstein unveiled his general theory of relativity and transformed our view of the fabric of the physical world. The theory, which explains gravitation as a deformation of space-time by mass, is a pinnacle of modern physics that underpins cosmology, from work on black holes to measurements of gravitational waves , and is used routinely to guide space missions and GPS satellites.
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in