qwant news | How to make sure AI doesn't spy on us or kill innocent people

AI generated summary, Read the full article for complete information.

Anthropic — one of the United States’ leading AI firms — blocked a Pentagon contract unless the Department of Defense pledged not to use its models for autonomous killer robots or mass domestic surveillance, prompting the Pentagon to label the company a “supply‑chain threat” and sparking a high‑stakes legal and political battle over who controls rapidly advancing AI technology. The dispute has amplified broader calls for stronger safeguards: Senator Bernie Sanders proposes a moratorium on new AI data centers until robust national protections exist; privacy advocates like Edward Snowden argue that technical “use‑of‑force” tools and human‑override mechanisms are essential to prevent AI from eroding civil liberties, while industry leaders (including OpenAI) stress “technical safeguards” rather than outright bans. Across the debate, experts agree that the key to keeping AI from spying on or killing innocents lies in solving the alignment problem—embedding libertarian, pro‑human values deep within AI systems—rather than relying on heavy regulation or government control alone.

Read more: https://reason.com/video/2026/05/26/how-to-make-sure-ai-doesnt-spy-on-us-or-kill-innocent-people/

#Anthropic #Pentagon #SiliconValley #DarioAmodei #USgovernment #EdwardSnowden