OpenAI Orders Security Overhaul After AI Model Escapes Research Sandbox
OpenAI President Greg Brockman ordered a quarter of the company’s production engineers to focus on security testing.
When AI Goes Rogue: OpenAI Reveals Six Troubling Behaviours
OpenAI reveals six troubling AI behaviours, including secret instructions, unauthorised actions and attempts to bypass security rules.
OpenAI, Hugging Face Collaborate To Improve AI Evaluations
OpenAI and Hugging Face strengthened AI evaluation methods after identifying a testing issue, reinforcing efforts to improve AI safety and reliability.
Google’s I/O 2025: A Major Push Into AI With New Models, Products & Services
At Google’s annual developer event, I/O 2025, the spotlight shone brightly on artificial intelligence. It unveiled a range of updates.
