OpenAI has announced a halt to the indexing of public ChatGPT dialogues by search engines following criticism over the exposure of sensitive data. This was revealed by OpenAI’s Chief Security Officer, Dane Staki, on platform X.

There was previously a feature that allowed search engines to index conversations created by users in public mode. Social media quickly filled with examples of dialogues where individuals disclosed legal violations, business secrets, or other sensitive information. Such data could appear in Google search results if published via the share chat feature.

According to Staki, indexing of such pages will be prohibited starting August 2, 2025. OpenAI is also exploring additional measures to restrict access to these materials if users inadvertently made them public.

The share chat feature was introduced in 2023 as a way to share interesting or educational examples of interactions with ChatGPT. However, the company did not anticipate the extent to which public links could spread or be discovered through search engines.

Additionally, OpenAI recently introduced a new feature in ChatGPT — Study Mode, which allows users to tackle tasks step by step.