| Today, State Senator James Maroney (D-Milford), Senate Chair of the General Law Committee, led the Senate’s passage of Senate Bill 2, landmark legislation and a primary Senate priority in the 2024 legislative session that provides landmark guidance and oversight of the use of artificial intelligence. The bill passed by a 24-12 tally, with Democrats in support and Republicans opposed.
While AI has existed for decades, the technology’s uses have rapidly expanded in recent years, posing opportunities but also pitfalls for rapid and significant adoption; this legislation seeks to provide guardrails for its use, allowing for innovation while also targeting AI adoption in ways that could provide detriment to the public.
“The advent of ChatGPT, OpenAI and Midjourney, to say nothing of TV commercials airing daily advertising AI’s latest developments, show that it’s AI’s world and we’re just living in it,” said Sen. Maroney. “It’s with that knowledge and context that we can’t wait to act. We saw, in the last 15 years, how light regulations impacted how social media and other emerging technologies transformed our world. They have benefitted us in many ways, but harmed society in others. This legislation seeks to place guiderails on AI’s uses and development. It’s focused on preventing harms while fostering innovation and benefits – and factors in the consequences of the technology’s future uses if we don’t act now.”
Senate Bill 2, “An Act Concerning Artificial Intelligence,” includes vital consumer, tenant, employee and citizen protection elements that strike a balance between unregulated use of AI and focusing on elements of its use most likely to have unjust consequences on consumers. The bill targets risks of AI discrimination based on race, age, religion, disability and other protected classes; it imposes duties of care on business only when AI’s utilization can have significant impacts, including in housing, finance, education enrollment, criminal justice, employment, government services or insurance.
One of the foremost elements of Senate Bill 2 protecting the public regards the use of “deepfakes,” which are manufactured images, videos or content using AI to simulate or manipulate individuals’ images and behaviors. This issue was prominently raised earlier this year when simulated, false pornographic images purporting to feature famous people, including Taylor Swift, circulated widely on the internet; they have also been used against average people, including teenagers, who have experienced serious distress due to its consequences. Senate Bill 2 would update current criminal statutes to include “deepfakes” in the consideration of unlawful dissemination of intimate images.
Additionally, with deepfakes’ use also growing in political advertising – nearly 80% of the country currently is considering measures to at least identify deepfake ads or calls, underscoring the growing issue – Senate Bill 2 would prohibit the distribution of AI-generated media attributing false words or actions to a person if the media is likely to disrupt electoral prospects of a candidate. This prohibition would represent a misdemeanor, with stronger penalties for those seeking to reach audiences of 10,000+.
The bill further ensures AI-created content published online will be tagged, disclosed or identifiable as AI providing users with the ability to evaluate what they read online.
Beginning in 2026, per the bill becoming law, developers would use reasonable care to protect consumers from risks of algorithmic discrimination and be required to provide information regarding limitations and potential benefits of such systems. AI system deployers will also be required to develop risk management policies and programs that specify and incorporate how the deployer will identify, document and eliminate risks of AI discrimination. Deployers will also need to regularly review their AI systems and their updates and changes to use. The Attorney General would have oversight to monitor such violations.
While the legislation seeks to monitor and prevent misuse and ill intent behind AI’s use, it also approaches the technology acknowledging its benefits. State agencies, as well as the Department of Administrative Services would be tasked to study ways that generative AI could be used to improve their processes and procedures; the state would develop workforce training programs with AI as a focal point; and the “Connecticut Citizens Academy” would be developed in state higher education to provide professional training on the use of the technology.
Before the bill’s passage by the State Senate today, it passed the General Law Committee unanimously on March 12 and the Judiciary Committee on April 22 by a 29-6 tally. It now heads to the House for further consideration. |
