AI Use by the US Government

The US government has acknowledged widespread AI use in 3,611 active or planned cases, a 70% increase from the previous administration. This raises concerns about handing over sensitive functions to AI, including life-and-death decision making. Examples include AI-assisted grant application screening, inmate misconduct assessment, and nuclear reactor control. While AI can be implemented responsibly, concerns about policy alignment and accountability persist. Engineers should be aware of these developments and consider their implications.