From the Research Universities edition of November 2, 2026
Preprint: a coding agent chose its own research questions in five pre-registered runs, all producing re-computable papers rated near workshop level, with low novelty
Ko Fujimura, Otsuma Women's University (single-author arXiv preprint) · Academic research on autonomous AI research agents · Japan
- Publisher
- arXiv
- Original publication
- Submitted 29 October 2026 (arXiv v1)
- Source retrieved
- 2026-11-03
What happened
An off-the-shelf coding agent, managed by Claude Code with Claude Opus 5.5, was given a roughly 7,300-word instruction document, a question-formulation skill (with one critique from gpt-6-astra under a US$20 budget) and a peer-review skill in which a reviewer agent reruns the study from its records. After 13 development runs, the system was frozen and five consecutive runs (one pilot counted by disclosed amendment, four pre-registered) were all reported. All five went from question to a revised, re-computable paper on a laptop with no human research input; external AI reviewers rated them at or near workshop level, but novelty and generality remained low. Small wording changes to the research-area setting shifted topics substantially. Records include claims-to-evidence logs, API costs, Git tags and a verification entry point.
Why it matters
Relevant to research offices setting policy on agent-generated papers: the study shows end-to-end agent research is feasible on modest hardware and that full logs and pre-registration make it auditable, while also showing low novelty and AI-only evaluation. It informs disclosure, authorship and review rules.
Evidence and measured results
Read the abstract, introduction, design, discussion and limitations. Five runs with one research model, no ablation, AI-only evaluation (human evaluation planned), in-loop reviewer from the same model family, and post-hoc choices disclosed as deviations. The paper itself was drafted by the management agent. Preprint, not peer reviewed.
Limitations and uncertainty
Very small sample, single author, AI-only ratings uncalibrated, no human evaluation yet; one research area setting tested. Non-US source; transfer to US institutional settings depends on local policy.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-11-02; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- Researchers can now generate whole papers with agents, and institutions lack rules for disclosure, records and authorship.
- Stakeholders
- Research integrity officer
- Graduate school dean
- Vice president for research
- Library scholarly communications
- Discovery questions
- Do current AI policies cover fully agent-run studies?
- What records must accompany AI-assisted manuscripts?
- Who reviews agent-produced work before submission?
- Value hypothesis
- A records standard may let institutions permit agent research while keeping it auditable.
- Potential engagement
- A review of agent-generated manuscript policy and required records package.
- Unsupported claims
- Agents produce publishable research
- AI reviewer scores reflect quality
- Results generalize beyond five runs
Pre-sales engineering
Role takeaway
- Fit
- Fits computational studies run by coding agents with full logging.
- Architecture and integration
- Separate management, research and reviewer agents; Git-tagged drafts; claims-evidence logs; AI-free verification entry point.
- Prerequisites
- Sandboxed workspaces
- Version control
- Run logging and cost records
- Constraints
- Isolation enforced only by instructions
- Same-family reviewer
- Computational topics only
- Security
- Enforce run isolation and data access with permissions, not prompts.
- Proof of value
- Independent reimplementation of a sample analysis
- Human expert rating of outputs
- Verification script regenerates all numbers
Delivery
Role takeaway
- Work
- Write disclosure policy, define records package, train reviewers and pilot with volunteer labs.
- Dependencies
- Existing AI use policy
- Repository and logging tools
- Faculty reviewers
- Ownership
- Research integrity office with graduate school and library.
- Skills and adoption
- Researchers learn pre-registration and records practices for agent work.
- Governance checkpoints
- Human review before submission
- Authorship and AI-involvement statement
- Pre-registration check
- Acceptance criteria
- Each agent-assisted manuscript has a verifiable records package (proposal)
- Disclosure rate tracked
- Reviewer training completed
- Risks
- Low-novelty paper flood
- Unverified AI evaluations
- Isolation failures between runs
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Management agent, research agent and reviewer agent in separate workspaces; local laptop computation; one budgeted API call to a different model for critique.
Governance
Who approves, reviews and stays accountable for outcomes?
Pre-registration, reporting every run and a claims-evidence record are concrete controls institutions could require for agent-produced work.
Security and privacy
What data, permissions and controls need testing?
Isolation between runs relied on configuration and instructions, not access control, a weak control for sensitive settings.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
Suggests the scarce human skill is choosing questions that matter, which the agent did poorly.
Procurement
What should contracts, pricing and exit terms secure?
Uses commercial model subscriptions and small API budgets; no special infrastructure.
Operating model
Which teams own the service once it runs?
Agent-run studies need records and review built in; reviewers rerunning code is not independent implementation.
Publication history
- 2026-11-02Research Universities · Issue 553 resources
Stable resource ID: fujimura-autonomous-research-agents-formulate-own-questions-20261029