Artificial intelligence is becoming a test of who governs Australia – elected representatives or technology companies with the resources to set the pace and terms of change. The central question is far bigger than whether an AI system follows instructions; it’s whether the corporations building these systems should also decide which risks are acceptable, when governments learn about them and who carries the consequences.
The scrutiny of OpenAI, Anthropic, Microsoft, Google and DeepMind at Australia’s parliamentary inquiry into AI brings that imbalance into focus. Questioning executives about AI safety and accountability matters, but a hearing cannot substitute for enforceable obligations. When governments depend on companies to explain and evaluate their own technology, corporate disclosure becomes a gatekeeper of public oversight. Australians are left being asked to trust an industry whose commercial incentives reward rapid deployment.
The deeper political challenge is employment. As AI moves into education, journalism, public administration and professional services, the promise of greater productivity raises an unavoidable distributional question: who receives the gains? Faster decisions and lower operating costs may benefit employers without producing better wages, stronger public services or secure work. Workers displaced by automation cannot pay the mortgage with assurances that technological progress will eventually create something better.
Australia’s biggest problem might be that it focuses on regulating tech industry behaviour while leaving behind the most consequential social effects that these companies are creating.
