The state of Connecticut is taking strides towards regulating AI systems with the introduction of Senate Bill 2, titled “An Act Concerning Artificial Intelligence.” The bill seeks to establish clear regulations and standards for the development, deployment, and use of specific AI systems within the state. Here are the key provisions of the bill:

• Requirements for Developers of High-Risk AI Systems (Sections 2-3): Developers of high-risk AI systems, those influencing critical decisions, must disclose limitations, risks of algorithmic discrimination, training data, and mitigation measures. They are also required to conduct impact assessments and maintain records.

• Requirements for Developers of Generative AI Systems (Section 4): Developers of generative AI systems, which produce synthetic content, must mitigate risks of unfair practices, discrimination, and privacy violations. Similar to high-risk systems, they must complete impact assessments and maintain records.

• Disclosure Requirements for AI Systems Interacting with Consumers (Section 5): AI systems interacting with consumers must disclose their nature unless evident or inaccessible to consumers directly.

• Exemptions (Section 6): Exemptions are provided for legal compliance, law enforcement cooperation, safety, research, and assisting with AI obligations.

• Enforcement (Section 7): The Attorney General is granted exclusive authority for enforcement, with provisions for a cure period and presumption of reasonable care for compliance.

• Artificial Intelligence Advisory Council (Section 8): An advisory council will engage stakeholders, make ethical AI recommendations for state government, assess the AI Bill of Rights, and suggest legislation.

• Unlawful Dissemination of Intimate Images (Section 9): The definition of “image” is expanded to include synthetic images for unlawful dissemination cases.

• Deceptive Media in Elections (Sections 10-11): Prohibition on distributing deceptive media in elections unless accompanied by proper disclaimers.

• State Agency Study and Pilot Projects (Section 12): State agencies are tasked with studying generative AI’s efficiencies, proposing pilot projects, and soliciting employee input.

• DAS Inventory and Assessments (Section 13): The Department of Administrative Services conducts inventories and assessments of AI systems used by state agencies to prevent discrimination.

• Chief Workforce Officer Duties (Section 14): Incorporation of AI training into workforce programs and guidance related to AI.

• Connecticut Citizens AI Academy and Certificate Programs (Sections 15-16): Establishment of an AI academy and certificate programs by the Board of Regents for Higher Education.

• Economic Development Initiatives (Section 17): Development of high-performance computing services and a CT AI Symposium for AI business growth.

The bill aims to establish a comprehensive framework for responsible AI development, addressing transparency, fairness, privacy, and ethical concerns. It marks a significant step towards accountable and transparent AI practices in Connecticut. The bill most recently had a public hearing after being referred to committee.

If you’re wondering how Connecticut’s AI bill, or any other bill around the world, could impact you, don’t hesitate to reach out to BABL AI. Their Audit Experts are ready to provide valuable assistance while answering your questions and concerns.