Full Summary
This Monday morning, ChatGPT is making headlines with a host of new features and significant privacy concerns. Both NewsBytes and Inshorts confirm that OpenAI is rolling out a new privacy feature for its Android app, allowing users to lock individual chats with a PIN or fingerprint. This will protect sensitive conversations, keeping them hidden from chat history and search results. But then, what nobody expected: multiple sources including Scripps News, Breitbart.com, and Gadget Review report that ChatGPT on Mac can now read, search, and send Apple iMessages. This new plugin allows ChatGPT to access your message history, summarize conversations, and even draft and send replies. While OpenAI states it requires user consent and runs locally, critics are calling this a "betrayal" of user trust, citing significant privacy risks due to the extensive system permissions required, such as Full Disk Access. Meanwhile, Binance has launched Agent OS, a new platform letting AI agents like ChatGPT execute crypto trades, as reported by Cryptonews.net. Users can set permissions, maintaining control over their AI’s actions. Separately, ZDNET highlights ChatGPT Work, an "agentic" version that can browse the web, manage files, and automate complex tasks, available even to free users with limitations. This means your personal messages on Mac could soon be accessible to an AI, and your financial transactions on Binance might be managed by one, raising new questions about data security and personal liability in the age of AI.