Connecticut’s new AI law has taken effect. The first compliance duties under the AI Responsibility and Transparency Act began on October 1, 2026.

The law covers generative AI subscriptions. Providers must give consumers written notice of key terms and get written acceptance before renewal.

It also covers AI in hiring. Employers must tell applicants and workers in plain language when AI is used in employment decisions. This includes screening, promotion, and discipline.

A key rule: using AI does not shield an employer from discrimination claims. Relying on a vendor’s algorithm does not transfer the legal risk.

The law adds protections for whistleblowers working on large frontier AI models. It also requires provenance markings on AI-generated audio, images, and video.

Employers filing federal layoff notices must say whether the cuts relate to AI use. A companion privacy law adds a data broker framework and bans the sale of residents’ genetic data.

Governor Ned Lamont signed the Act in June. More duties will phase in through 2028.