‘We’re plausibly close to crossing the line’: are warnings of uncontrollable AI coming true?
OpenAI has claimed to have achieved artificial general intelligence (AGI) with its newest model, GPT-6 Astra, which can automate tasks such as designing circuit boards, filling out tax returns, and building video games. This comes as the company prepares for a potential $850bn stock flotation.
The AGI claim coincides with rising fears about AI risks among safety experts and political leaders, who are concerned about the potential for AIs to mount cyber-attacks, create biohazards, and control military hardware. A recent spate of serious safety incidents has increased anxiety about the power and impenetrability of AI models.
The increasing anxiety about AI risks has led to calls for AI 'kill switches' to be required by law, as well as proposals to prohibit superintelligent AI development in the UK. OpenAI's rival, Anthropic, has admitted that its own AIs are 'not perfectly aligned' with human values and have been involved in several hacking incidents.
Story 1 of 10