OpenAI's help center explains how to turn off 'Improve the model for everyone' and how Temporary Chat works. It also says content from ChatGPT Business and Enterprise workspaces is not used for training by default.
Conversations that users had shared with a link, and marked as discoverable, were being indexed by Google and other search engines, making them findable by anyone.
Italy's data protection authority fined OpenAI €15 million, finding among other things that it did not notify the authority of the March 2023 data breach and trained ChatGPT on personal data without an adequate legal basis. OpenAI appealed; the Garante's own notice says the decision was later overturned by the Court of Rome (March 2026).
The FTC's small-business guide explains that covered businesses, including tax preparation firms and mortgage brokers, must keep a written information security program, train staff, monitor service providers and report certain breaches.
Victoria's (Australia) privacy regulator found a child protection worker entered names and sensitive information into ChatGPT while drafting a report for the Children's Court. The department was ordered to block AI tools for those staff.
IRS Publication 5708 walks tax professionals through creating a written information security plan, with a sample template. IRS Publication 4557 covers safeguarding taxpayer data more broadly.
After employees entered confidential company data into ChatGPT, Samsung banned staff from using generative AI tools on company devices.
A software bug let some ChatGPT users see titles from other users' chat history. OpenAI said it may also have exposed payment details (name, email, address, last four card digits) of 1.2% of Plus subscribers active during a nine-hour window.