AI writes 214-page wiki from 45,047 Enron emails, each fact cited
The team gave an AI 45,047 Enron emails and it produced a 214‑page wiki covering the companies, regulators and banks Enron dealt with. Every sentence includes a citation to the originating email, and facts that later became false are shown as expired on the date they changed. A second AI performed a held‑out audit to verify the pages.
Key points
- AI generated a 214‑page wiki from 45,047 Enron emails, citing each fact to its source email
- Pages cover entities mentioned in at least 25 messages and five key Enron parts
- The dataset is public thanks to FERC release and Carnegie Mellon maintenance since 2003
The wiki groups entities into pages for organizations mentioned in at least 25 messages and for the five most‑discussed parts of Enron itself – three business units, a trading platform and a trading system. In total the system identified 209 pages with 929 entities and 5 pages with 551 entities, covering 57 narrow types, 78 products and platforms, 76 facilities and market points, and 49 events. The source data is public because the Federal Energy Regulatory Commission (FERC) released the emails during its investigation, and Carnegie Mellon has maintained the archive since 2003.
The authors argue that such a “company brain” sits beside traditional CRM systems, holding knowledge that changes too often for static fields, and they provide the prototype as a reference for anyone wanting to build similar systems.
Show HN: We had AI write a wiki from 45k Enron emails, every fact cited
thisandthat.chat · 23 October 2026Loading the full article…
This text was published by thisandthat.chat. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
Coverage and discussion
1source- Hacker News discussion · 43 pointsnews.ycombinator.com
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Agents & Tools
All →- OpenAI, PerturbAI, and Scripps Research used thousands of AI agents to analyze mouse brain data in 32 minutes · 6 src
- Meta pauses Muse AI sharing after researchers find 6,800+ chats in public DNS records · 8 src
- Five free AI coding agents you can try in 2026 · 1 src
- OpenAI Academy offers free AI courses and certification, reviewer says · 1 src
- Schneider Electric, NVIDIA and HPE launch industrial AI agents for edge operations · 2 src
Comments
via GitHub Discussions