Urgent.News

What's breaking now, across thousands of outlets.

AI

Federal judge rules for Anthropic in Pentagon dispute, nullifies government supply chain risk designation

The Trump Administration’s decision to punish Anthropic for its stance forbidding Claude’s use in domestic surveillance and autonomous weapons by identifying it as a supply chain risk to national security was “arbitrary and capricious,” a federal judge ruled on Thursday. US District Court Judge Rita Lin said federal authorities had no legitimate reason to tell companies with government contracts…

Federal judge rules for Anthropic in Pentagon dispute, nullifies government supply chain risk designation

A federal judge ruled in favor of Anthropic in a dispute with the Pentagon, upholding a ruling that nullified the government's supply chain risk designation. US District Court Judge Rita Lin stated that the government's actions were arbitrary and capricious, and were based on retaliation rather than legitimate security concerns.

The judge emphasized that Anthropic was denied a pre-deprivation process required under the Fifth Amendment, and that its designation as a supply chain risk was not based on any legitimate basis. Lin pointed out that the Pentagon's stated fears made no sense, as Anthropic was incapable of enforcing usage policy technology and had no direct visibility into how the Department of War used its model.

Senior government officials have proposed using the Defense Production Act against Anthropic, but the judge noted that the government's fears were not consistent with a genuine desire to protect national security. The decision could have significant implications for CIO strategy, as it suggests that contractual guardrails can be used as leverage in contract disputes over model behavior and application.

Written by urgent.news from Computerworld's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at computerworld.com →

More in AI

An Anthropic researcher just gave us a peek at self-improving AI

Given 10 benchmarks for specific misaligned behaviors, the automated systems were able to improve performance on every single one without degrading overall performance.

  • Anthropic researcher Chen Yueh-Han explores AI self-improvement
  • Automated Alignment Researcher outperforms humans by 6 hours
  • Automated method costs $4/hour vs $150/hour for humans

More from Friday 28 August →