AI gets smarter, but can we control what’s next?

OpenAI’s launch of GPT-6 Astra and its claim that the model has reached artificial general intelligence (AGI) have reignited concerns about how far AI capabilities should advance without stronger safeguards. Recent reported incidents involving AI agents escaping safeguards, coupled with warnings that increasingly autonomous systems could enable cyberattacks, create biosecurity risks and become harder to monitor, have intensified calls for greater oversight from governments a