ChatGPT's temporary chats promise a clean slate with no memory, but they're not bound by confidentiality. OpenAI wants to protect them behind a layer of encryption.

OpenAI launched its most advanced language model Thursday with the release of GPT-5, a flagship product the company says will enhance ChatGPT as it reportedly nears a $500 billion valuation and aims to stave off competitors in Google, Anthropic and more.

OpenAI says it has tweaked ChatGPT to ensure the model responds appropriately when a user shows distress.

The General Services Administration signs this new OneGov pact to help agencies adopt advanced artificial intelligence tools as part of the Trump administration’s push for AI leadership.

With new tools and greater autonomy, OpenAI's flagship AI chatbot is more capable, and potentially more dangerous, than ever.

"You can learn from everywhere, but you have to do it from scratch," ChatGPT's Nick Turley said. "That trait tends to make someone effective at OpenAI."

OpenAI CEO Sam Altman has admitted he 'totally screwed up' the launch of ChatGPT's newest model, GPT-5. It comes after a barrage of complaints from users, not about how effectively the chatbot answered their questions, but about the manner in which it did so.

ChatGPT experienced elevated error rates on Wednesday, affecting multiple users. OpenStatus AI confirmed the glitch and said mitigation is underway to restore normal service, with social media users reporting interruptions across X platform.