The study from the Brookings Institution and the Centre for the Governance of AI is concerning, but there are proactive steps ...