The AI ‘Slowdown’ Is an Antitrust Mess
wired ·
Leading AI companies have called for a coordinated development “slowdown” amid reports of dangerous AI agents and growing concerns about model safety. Antitrust experts warn that describing the effort as a slowdown or pause could make it look like an agreement to reduce output, even if the goal is to prevent catastrophic risks. They say the companies could have framed the initiative as cooperation on safety standards, which might increase output and competition under existing legal doctrines. Experts also note that industries can reduce antitrust risks by creating standards organizations and notifying federal regulators, though exemptions can draw criticism for strengthening dominant firms. The major labs face pressure from investors, competitors, and political leaders to keep developing quickly, while some employees argue that market incentives are undermining safety. If regulators investigate, the probe could last years and impose extensive document, deposition, and litigation burdens, leaving the companies to develop safety rules largely on their own.
wired