The year is 2025. Catching the wave of excitement over its possibilities, many government organizations began rolling out generative AI capabilities for a range of applications, and many of those capabilities are now in place. However, something horrible happened. One of the systems has 'hallucinated', producing catastrophically incorrect content that has led to a spectacular failure. It makes national news, and a congressional investigation follows. In the fallout, the DoD adopts a policy banning the use of GenAI, and ongoing AI programs are paused indefinitely. Luckily it is still 2024, and this is a future we can avoid. How?