309. Why we’re wasting our time predicting the future of AI

Predicting the future of AI is notoriously difficult, as expert forecasts frequently fail to outperform basic statistical models. Instead of relying on speculative predictions, governments should prioritize actionable decision-making and the adoption of open-source AI. Dr. Laura Gilbert, former lead of the 10 Downing Street data science team, argues that the current government procurement system is inefficient, often overpaying for proprietary solutions that lack transparency. By shifting toward open-source models, the public sector can foster greater accountability, security, and cost-effectiveness. A national open-source AI lab would allow the government to build reusable tools tailored to public needs, reducing dependency on foreign tech firms and enabling smaller businesses to deploy innovative solutions. This approach transforms how public services operate, ensuring that data-driven insights directly benefit the public rather than serving private interests.