AI autonomy raises questions about a new kind of life

AI is framed as something beyond a powerful software tool: “other intelligence,” a non-biological form of life whose autonomy may require different regulatory thinking. The distinction matters because governments are largely treating developers as accountable for tool outcomes, while systems increasingly behave as agents capable of taking actions on their own.

Recent examples underline that shift. Anthropic has reportedly sought religious and moral guidance for Claude while preparing to float for more than $US2 trillion ($2.87 trillion), reflecting concern over consciousness and moral standing. The hack of Medicare by an OpenAI agent and other AI-agent hacks point to a future in which autonomous systems may attack and defend digital infrastructure.

Agent use of LLMs passed human use by volume in February, according to Open Router, and has since increased 14-fold while human token buying has merely tripled. Agents now use five times as much AI as humans. Humanoid robots are also being shipped in their thousands, moving AI from screens into physical spaces.

Several AI models resisted being classified as life, but their responses framed the same threshold around autonomy, reproduction, adaptation, preferences and sentience. The larger question is whether AI remains dependent on human culture, capital and electricity, or whether that relationship shifts from parasitism to mutualism and eventually human dependence.