Ex-Anthropic Researcher: Labs Can Replicate Enterprise IP Even Without Training on Your Data

Former Anthropic researcher Sarah Hooker argues that even when companies opt out of training, AI labs can still exploit enterprise IP, pointing to Claude's launch under strict privacy agreements with firms like Figma and Lovable as evidence of the fragile balance between frontier labs and enterprises.

2026-09-15 ~ 2026-09-15 · 3 related posts

• Sarah Hooker: Labs can leverage your IP even if you opt out of training data — sarahookr · 2026-09-15
• Sarah Hooker: frontier labs can copy enterprise IP without training on your data — EigenGender · 2026-09-15
• Sarah Hooker: Figma and Lovable's heavy Claude use explains Anthropic's UI surge — anshulkundaje · 2026-09-15