A Hacker News user reports that OpenAI has repeatedly re-enabled the 'allow training' setting after they manually disabled it. The user documented the timing of their last reset and found the setting reactivated upon checking. This has sparked discussion about OpenAI's trustworthiness, with commenters citing the company's history of questionable practices including the 2023 CEO drama involving Sam Altman, allegations of IP theft from researchers like Levent Alpöge, and a perceived pattern of prioritizing growth over user privacy.

Some users report similar experiences with Anthropic's Claude, suggesting the issue may be broader. Others question the original poster's claims due to lack of proof, while many express distrust in major AI labs, leading clients to prefer on-premises or open-source models. The conversation highlights ongoing concerns about data consent and corporate accountability in AI.