Definition
Plain language
An AI lab known for systems that try to automate the whole research process.
As stated in the literature
Sakana AI, the lab behind the AI-Scientist autonomous-research pipeline (including AI-Scientist v2); used as a baseline in Chain-of-Evidence integrity audits, where its tree-search workflow exhibited reporting-integrity failures on a benchmark it wasn't designed for.
Why it matters: It matters because audits of such autonomous-research systems reveal how easily they can produce reporting-integrity failures worth guarding against.
For example, this lab's AI-Scientist pipeline attempts to run an entire study—forming a hypothesis, experimenting, and writing it up—with little human involvement.
Heard on the show
“Their own, ScientistOne, plus four open-source baselines — the most well-known is Sakana's AI-Scientist v2, the one that produced the workshop-accepted papers I mentioned earlier.”Episode 089 — When AI-Written Papers Read Well But the Evidence Underneath Is Broken