Opinion | I Left My Job at OpenAI. Putting Ads on ChatGPT Was the Last Straw.

Zoë Hitzig, a former researcher at OpenAI, resigned after the company began testing ads on ChatGPT, viewing it as the final confirmation that the company was abandoning its initial safety and ethical considerations. While acknowledging the need for revenue to run expensive AI models, Hitzig argues that advertising based on the candid, private data users share with ChatGPT creates an unprecedented potential for manipulation. She rejects the false dichotomy between restricting access to wealthy users or exploiting users via ads, proposing alternative funding structures. These include explicit cross-subsidies, where businesses using AI for high-value labor pay a surcharge to support free access, similar to utility infrastructure funding. Another option is pairing advertising with binding governance structures, featuring independent oversight boards with authority over data use. A third approach involves placing user data under independent control via a trust or cooperative that legally acts in users' interests. Hitzig fears that without these structural changes, the industry will default to either manipulating non-paying users or exclusively benefiting the wealthy who can afford high-cost subscriptions.

(Source：Nytimes)