We must also recognize that advanced AI presents serious risks. In the wrong hands, AI can be weaponized against the American people. More broadly, there are unresolved questions about whether we currently have the ability to keep powerful AI systems under our control. These questions are especially salient as companies pursue artificial superintelligence—AI models that can do anything a human can do but much, much better.
Jim Banks
U.S. senator from Indiana
In a May 2026 letter to senior Trump Administration officials, Jim Banks wrote that advanced AI presents serious risks, that in the wrong hands AI can be weaponized against the American people, and that there are unresolved questions about whether powerful AI systems can currently be kept under human control as companies pursue artificial superintelligence.
U.S. Senator for IndianaMay 13, 2026Letter to senior Trump Administration officials on advanced AI and national security
Context and checks
Surrounding words
AI systems with such research and development capabilities would potentially lead to rapid and self-reinforcing improvements: America must reach this threshold before the PRC. We must also recognize that advanced AI presents serious risks. In the wrong hands, AI can be weaponized against the American people. More broadly, there are unresolved questions about whether we currently have the ability to keep powerful AI systems under our control. These questions are especially salient as companies pursue artificial superintelligence—AI models that can do anything a human can do but much, much better. Our major AI companies believe superintelligence could pose significant risks that our current scientific understanding is not yet capable of addressing.
What this quote does not say
- The passage describes unresolved control questions and serious risks; it does not predict extinction or assign any probability.
How this was checked
- Quote located in the captured page; spacing normalised for display.
Reviewed by
an AI reviewer that read the whole source · an automatic character-by-character check