Economist Proposes AI-Reformed Peer Review: AI Report First, Then Human Referees
_akpiper · x · 2026-09-06
A forwarded proposal for reforming econ journal refereeing with frontier AI: rather than referees individually prompting ChatGPT, editors would first generate a frontier-AI report (via an open-source paper reviewer journals can modify), authors would then respond to and push back on the AI report's points, and only then would human review proceed — keeping human judgement central while using AI as a screening and communication tool.
Related event: NBER Paper: AI Review Reports Work Best Given to a Single Reviewer(2 posts)→
More from Research
- Georgetown's AI referee leaderboard has Claude Opus 4.8 scoring finance papers — emollick · 2026-09-07
- DeepMind's WeatherNext 3 Trains on Raw Satellite Data, Drops 6-Hour to Hourly Forecasts — dair_ai · 2026-09-07
- HF engineer trains Qwen3.5-9B with GRPO and GEPA to code Blender 3D rooms, GLM-5.3-Flash as judge — mervenoyann · 2026-09-07
- First-in-human CRISPR therapy safely and continuously lowers cholesterol — garrytan · 2026-09-07
- Proposal: add chess RL environments to LLM training runs, with and without Stockfish — burny_tech · 2026-09-07
- Astra architecture debate: recurrent depth vs temporal recurrence, and a strange OpenAI leak — burny_tech · 2026-09-07