OpenAI fires 3 researchers, scraps GPT-6.1 Astra, delays IPO
Technology
OpenAI just let go of three researchers for leaking confidential information to an artificial intelligence (AI) safety group, breaking the company's trust.
This follows other recent security slip-ups with AI systems.
Because of these incidents, OpenAI has scrapped plans to launch its AI model GPT-6.1 Astra over its safety concerns and pushed back its IPO plans.
OpenAI tightens testing rules, increases transparency
To tackle these issues, OpenAI has rolled out a new monitoring system to catch problems faster and made security rules stricter for engineers testing AI.
They're also being more open about when their AI models mess up, all part of a bigger push to keep things safe and transparent as calls for better regulation grow louder.