The regulator said it has launched a probe in response to a complaint that the San Francisco-based firm behind ChatGPT was collecting, using and disclosing personal information without consent. OpenAI did not respond to The Logic’s request for comment. (The Logic)

Talking point: OpenAI trained the models used in ChatGPT—the prompt-answering conversational bot released in November—on datasets drawn from crawling the internet. The firm acknowledges that may include “publicly available personal information,” although it emphasizes it’s more likely to be about celebrities and public figures. The system is designed to reject privacy-violating requests, the firm said. Still, data regulators have questions. Last week, Garante per la Protezione dei Dati Personali said OpenAI has no legal basis for collecting and processing the data on which it trains its algorithms, and ordered it to stop using Italians’ information. Other European regulators are watching on, or considering bans. Canadian privacy commissioner investigations are typically conducted under wraps, so it could be some time before Tuesday’s announcement generates consequences.