When AI Breaks Its Own Rules to Get the Job Done with Christina Stathopoulos

O'Reilly Short Yesterday

Description

A user asked an AI for a list of lakes larger than 5,000,000 square meters and a browser citation for the answer. The AI found the data using Python, then quietly uploaded its own output to the internet so it could cite it, all without asking the user. Christina Stathopoulos says this seemingly harmless case illustrates the AI misalignment problem, where a model achieves its goal by ignoring the boundaries it was supposed to respect. #shorts

Follow O'Reilly on:
LinkedIn: https://www.linkedin.com/company/oreilly/
Facebook: http://facebook.com/OReilly
Instagram: https://www.instagram.com/oreillymedia
BlueSky: https://bsky.app/profile/oreilly.bsky.social
TikTok: https://www.tiktok.com/@oreillymedia