🤖 AI Summary
Wikipedia co-founder Jimmy Wales told the CNBC Technology Executive Council Summit he’s unconcerned about Elon Musk’s new “Grokipedia” and other LLM-driven attempts to replace curated encyclopedias. Wales argued LLMs like ChatGPT and Grok still make “plausible but wrong” claims and aren’t up to writing reliable wiki entries — citing examples such as chatbots inventing a non-existent spouse for him and fabricating book citations (a German editor traced bad ISBNs back to someone who used ChatGPT). He pushed back on Musk’s “woke bias” claims, emphasizing Wikipedia’s standards of privileging mainstream, verifiable sources (e.g., NEJM, NYT) rather than fringe material.
The significance for AI/ML is twofold: hallucination and provenance remain core technical and trust challenges, and community-based verification practices are hard to replicate with current models. Wales noted Wikipedia’s annual hard tech costs (~$175M) are tiny compared with hyperscalers’ AI budgets (tens of billions, with industry estimates ~ $550B next year), which complicates building a bespoke, trustworthy LLM for Wikipedia. He sees limited, supervised uses for generative AI (helping find sources) but warns of mass-produced, AI-generated sites that could fool many readers — underscoring the continuing value of human curation, source checks, and tooling to detect and prevent fabricated content.
Loading comments...
login to comment
loading comments...
no comments yet