Should AI companies be able to restrict how the US military uses their technologies?
Yes, said 44 percent of Americans, according to a poll by YouGov for the Economist.
By comparison, a quarter said the military should be able to use AI however it wants, while 32 percent were unsure, highlighting widespread public uncertainty over how the technology should be deployed in defense operations.
Skepticism Toward AI
The findings reflect broader public skepticism about artificial intelligence.
Nearly half of Americans expect AI to have mostly or entirely negative effects on society, while a much smaller share expect positive outcomes.
The poll also found Americans divided on whether military AI would improve national security: 22 percent said it would make the US safer, 25 percent said it would make the US less safe, and 38 percent were unsure. The remaining 15 percent believe it won’t make a difference.
Experience with AI appears to shape perceptions of the technology. The survey found 22 percent of Americans regularly use AI tools and 43 percent use them occasionally, while those with little experience are more likely to expect negative impacts.
Consumer Backlash
Public unease about military AI is also reflected in consumer reactions to defense partnerships.
After OpenAI announced a deal with the Pentagon, ChatGPT mobile app uninstallations surged 295 percent in a consumer backlash that prompted the company to revise its safeguards in the agreement.
Industry Pushback
Concerns about AI use in the military are also surfacing within the technology industry.
Google employees have urged company leaders to set limits on how its AI systems can be used in defense programs, warning that such systems could enable mass surveillance or autonomous weapons without human oversight.
Similar tensions have emerged at Anthropic, which has pushed for guardrails on how its models are used in projects linked to the US defense department.
