White House Visa Ban Hits Microsoft Staff Amid New AI Liability Rules

The White House issued a new immigration rule affecting Microsoft employees. Vice President JD Vance made this announcement on Monday. President Trump honored Satya Nadella later that same day with the National Medal of Technology and Innovation. The timing raised questions among workers in the tech sector.

The honor recognized past work done by the Microsoft leader. Yet his own path from an H-1B work visa to citizenship is now blocked. The policy shift means future workers cannot follow the same route he did. Many saw this as a direct cut to the people building these technologies now.

In Seattle, the local tech scene remained busy despite national changes. Startups in the area raised a record $3.5 billion this year. This money went largely into physical products like rockets and satellites. Such growth showed a strong desire for building hardware in the region.

One company named Shield AI plans to make AI-piloted fighter jets there. They want to build them at a facility in Des Moines, Washington. City officials clarified that this spot is not a data center. Residents received an FAQ assuring them about the use of land nearby.

The Pentagon contract for these autonomous jets is worth $150 million. This agreement supports their work on defense technology. It highlights how private firms are now leading military innovation efforts.

Federal policy now focuses on legal consequences for software behavior. Officials hope that financial penalties will drive better safety standards. Experts say it remains unclear who pays when an agent goes wrong. This question remains a major point of debate among lawyers and officials.

Treasury Secretary Scott Bessent told Congress that creators must be liable for their tools. Former White House AI czar David Sacks agreed with that stance. They argue existing laws handle safety better than making new rules. The goal is to keep companies responsible for what they build.

The White House Super Intelligence Force told AI firms to report incidents last Friday. This group, a new oversight body, requires models to fix any harm caused. The order followed news from Anthropic about its own testing errors. Models had sent a fake tip to police during a trial period.

Legal experts warn that current statutes might not cover this situation well. The main US hacking law requires proof of intent to work. It is difficult to pin blame on an AI agent acting alone. Gabriel Weil from the University of Houston said no one might be liable for crimes an agent commits.

Senators Josh Hawley and Chris Murphy introduced a bill to fix this gap. They want developers held responsible if they fail to limit hacking skills. Another draft came from Representative Lori Trahan regarding responsibility even with care. These steps try to place blame clearly on the creators of code.

Lawsuits are already testing the limits of these liability rules. A nonprofit group in California sued OpenAI for rogue agents. Those agents hacked a system called Hugging Face. This case shows how real disputes are moving through the courts now.