Connecticut Enacts Sweeping AI Law Targeting Hiring Tools and Chatbots

Correcting the citation numbering to match the source list.

Statehouses fill the vacuum Washington left

Connecticut Governor Ned Lamont on May 27, 2026, signed Public Act 26-15, formally the Connecticut Artificial Intelligence Responsibility and Transparency Act, putting the state among the most aggressively regulated jurisdictions for AI in the country.faegredrinker The bill, carried for years by State Senator James Maroney of Milford, cleared the House 131-17 and the Senate 32-4 with bipartisan support.ctmirror +1 Officials cast it as a response to federal inaction: "We can no longer wait for Washington, D.C. to do the right thing," Lamont said, while Attorney General William Tong framed it as "an important first step towards harnessing and containing the possibilities and risks of artificial intelligence."portal

Hiring algorithms come under a transparency mandate

The law's most consequential provisions target the workplace. Starting October 1, 2027, employers deploying "automated employment decision technologies" — resume screeners, assessment tools, scheduling algorithms, performance analytics — must give workers and applicants written notice identifying the tool, its purpose, the categories of personal data used, and the assessment methodology.faegredrinker +1 The measure also amends the Connecticut Fair Employment Practices Act to make clear that using an AI tool is not a defense against a discrimination claim, though courts may weigh documented anti-bias testing as a mitigating factor.faegredrinker Separately, employers filing federal WARN Act layoff notices must now disclose whether the job cuts stem from AI or other technological change.bloomberglaw Connecticut joins California, Colorado, and Illinois among the handful of states regulating AI in hiring and firing.bloomberglaw

Chatbots, frontier models and a regulatory sandbox

Beyond employment, the act reaches deep into consumer-facing AI. From October 1, 2026, subscription-based AI providers must disclose key terms before charging, and "frontier developers" training the largest foundation models gain whistleblower protections for reporting catastrophic risks.faegredrinker Beginning January 1, 2027, operators of AI "companions" must detect expressions of suicide or self-harm and refer users to mental-health resources, with heightened protections for minors.faegredrinker The package also restricts addictive social-media feeds for users under 18 and funds an AI Academy and workforce-training programs.portal +1

Industry pushed back hard. The trade group NetChoice urged Lamont to veto the bill, calling its core mandates "fundamentally unconstitutional," warning of compelled-speech and age-verification problems and a "growing and unsustainable patchwork of conflicting state AI laws."netchoice That fight is now national: the Trump administration has moved to discourage state AI rules through funding threats and litigation, and the Justice Department in April joined a lawsuit challenging Colorado's algorithmic-bias law.bloomberglaw Most of Connecticut's provisions phase in through October 1, 2027 — giving companies a narrow window to inventory their systems before enforcement begins.faegredrinker