FARS: A Fully Automated Research System Deployed at Scale

Summary

The largest public deployment of a fully automated AI-for-AI research system to date: FARS autonomously ran ideation, planning, experimentation and writing through stage-specific agents on a shared workspace and produced 166 complete papers across 67 fine-grained AI/ML topics, deliberately preserving intermediate artefacts as an auditable corpus rather than a curated highlight reel. Evaluated with 282 structured volunteer reviews over 140 papers including integrity checks and LLM-use disclosure; some outputs are review-worthy, but narrow experimental scope, methodological limits and integrity issues recur. Note: FARS is the reference system in the 'Can AI Evaluate AI Scientists?' benchmark.

Contribution

Queued — not yet read.

Method

Queued — not yet read.

Relevance to RISE

Surfaced by the 2026-08-29 literature sweep as a high-priority candidate for the RISE knowledge base (angle: autonomous AI scientists and their evaluation). Verified against: arXiv abstract page opened 2026-08-29 (title, ~55 authors, v1 30 Jun 2026 / v2 13 Jul 2026, full abstract read)

Critique / open questions

Queued — not yet assessed.

Key quotes

Recent automated research systems show that language-model agents can generate hypotheses, run experiments, and write complete manuscripts, but most evidence still comes from selected examples, human-framed topics, or a few pre-defined research tasks. (abstract)

We present FARS (Fully Automated Research System), a fully automated AI-for-AI research system designed to operate across research topics at scale. (abstract)