If the AI bubble does burst, taxpayers could end up with the bill (phys.org)

The Conversation warns that rapid, high-stakes investment in AI — driven by billionaire optimism about reaching AGI and by governments eager to deploy AI across public services — could leave taxpayers footing the bill if an AI market bubble bursts. Big AI firms now have combined market values exceeding £2 trillion and are tightly entangled through deals and investments worth hundreds of billions. At the same time, governments (the UK has declared it’s “all in”) are embedding AI into health, education, defense, courts and passport systems, making these firms’ services critical to everyday life. The warning is sharpened by a study showing about 95% of generative-AI pilots fail, highlighting a gap between hype and reliable production capability. For the AI/ML community this matters because systemic dependence on a few providers creates concentration risk, moral hazard and potential fiscal exposure similar to the 2008 bank bailouts. If an essential AI provider falters, public services, pension funds and government budgets could be forced to intervene. Key technical implications include the need for robust reliability, auditability and resilience in deployed models, and for policy measures that limit entanglement and protect taxpayers — such as transparency, contingency planning, and regulatory oversight of public-sector AI procurement.