Between September 11–16, 2026, Washington’s AI-control debate moved from abstract risk language into interrupt language that operators can recognize. Reuters reported on September 11 that Senate negotiators were debating a duty of care for AI…