OpenAI has reported that 53 user images from ChatGPT were posted on third-party services due to the way AI agents operate. The company is continuing its investigation and recommends measures to protect user data.
OpenAI has disabled the automatic selection of reasoning depth in ChatGPT Plus and Pro—now users must manually set their preferred level. Quick responses will no longer become more detailed without manual adjustment, and access to the highest levels is restricted based on your subscription plan.
Andreas Thom has expressed doubts about the transparency of OpenAI's use of user data following the similarity of methods used to solve a complex problem. He is urging the company to disclose more details, but OpenAI has not made any public comments on the situation.
The European Commission has tightened regulations on ChatGPT, Reddit, and Roblox, officially designating them as very large platforms under the DSA. These services are now required to meet additional obligations aimed at reducing risks and protecting users.
ChatGPT now allows you to connect multiple Google accounts at the same time, making it possible to manage your emails, calendars, and contacts from a single window. This new feature is available on paid plans and streamlines the management of both personal and work tasks.
The ChatGPT Atlas browser will be discontinued on August 9, 2026, and its main features will be integrated into the new ChatGPT desktop application. Users are advised to export their data in advance and avoid using Atlas after support ends due to potential security risks.
A new study has shown that modern AI systems can outperform the average person in creativity tests. However, the most creative individuals still remain unmatched. AI is becoming a powerful tool to enhance human imagination, rather than replace it.