ChatGPT lead on shifting from "can do" to "co-responsible" AI colleagues

xiaohu · x · 2026-09-02

Tara Seshan, product lead for ChatGPT Work and Codex, discusses collaboration as agents become persistent colleagues. With AI working in the background with project context, product design and org structures must evolve. Seshan frames the shift from models, environments, permissions, and interfaces to human responsibility, explaining how AI moves from task execution to long-term co-responsibility, while human roles shift to direction, taste, communication, and accountability.

Related event: OpenAI Product Lead on Designing for Persistent AI Coworkers(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →