Connecticut AI Rules: Injury Claims Shift in 2026

Connecticut is getting serious about AI regulation, and its new laws will directly affect consumer rights and how injury claims are handled. The state’s Public Act 24-119 is a major move that changes how artificial intelligence systems are governed, creating a new playbook for liability and giving real recourse to people harmed by AI. This framework sets clear rules for AI developers and deployers to ensure accountability and transparency, which creates new ways for individuals to seek compensation for AI-related injuries in Connecticut.

Public Act 24-119: The Foundation of Connecticut’s AI Regulation

Effective July 1, 2026, Connecticut’s Public Act 24-119, “An Act Concerning Artificial Intelligence,” creates a binding legal framework for the responsible development and deployment of AI systems. This isn’t a set of loose recommendations. The statute’s main goal is to get ahead of the risks that come with AI, especially around bias, discrimination, and privacy violations that can cause real-world injuries. It assigns specific duties to both AI developers and deployers, establishing a new standard of care that didn’t exist before.

The law gets its teeth from its definition of “high-risk artificial intelligence systems.” These are the AI tools that make consequential decisions about people’s lives, things like employment screening, credit applications, housing, insurance, and critical public services, and they are now under intense scrutiny. The act requires developers and deployers of these systems to perform detailed impact assessments to find and fix potential negative effects. A company’s failure to run these assessments properly, or to act on the risks they find, could become the core of an injury claim, which completely changes the game from the old days where proving an AI caused specific harm was an almost impossible uphill battle for a plaintiff.

The legislature is trying to prevent harm before it happens, but it’s also creating a clear path for legal action when things go wrong. Attorney General William Tong has been a major advocate for this law, stressing the need for real consumer protections as technology gallops ahead. When the state’s top lawyer says his office will actively enforce PA 24-119, businesses using AI in Connecticut need to take that very seriously.

Who Is Affected by the New Regulations?

Public Act 24-119 targets everyone in the AI lifecycle, but it draws a clear line between two groups: AI developers and AI deployers. Developers are the ones building or significantly changing the AI models. Deployers are the companies or people actually using the AI in their day-to-day operations. This distinction matters because the law assigns different, though sometimes overlapping, responsibilities to each, and you need to know which hat you’re wearing.

For instance, a software company that builds an AI algorithm for insurance companies to price risk is a developer. They have to design it with safeguards against bias and run it through impact assessments. The insurance company that buys and uses that algorithm to set policy premiums is a deployer, and they are also on the hook for using it responsibly and transparently. This dual responsibility means when an AI makes a bad decision that harms someone, the blame, and the liability, can easily fall on both of them, depending on where the failure occurred.

Beyond the tech companies, these rules reach into a huge number of sectors. Healthcare providers using AI for diagnostics, banks using it for loans, HR departments using it to screen resumes, and even state agencies are all affected. If your organization uses AI to make decisions that impact people’s lives or finances in Connecticut, whether you’re a startup in Stamford or a big corporation in Hartford, these new standards apply to you.

Concrete Steps for Businesses and Individuals

If you’re a business in Connecticut, you need to start getting ready for this now. That July 1, 2026, deadline will be here before you know it, and getting compliant is a big job that will require you to change some internal processes. First, you have to audit every AI system you’re using or building to figure out if it qualifies as “high-risk.” This means taking a hard look at the system’s purpose, the data it’s fed, and its potential impact on people.

Second, for any system you identify as high-risk, you must begin conducting AI impact assessments. These aren’t a one-and-done checkbox. They demand ongoing monitoring and updates as the system evolves. You’ll need crystal-clear documentation of the assessment, detailing the AI’s purpose, the data used, potential bias risks, and your strategies to fix them. I’m telling clients to put a specific person or team in charge of AI compliance. The consequences are too high to hand this off to a summer intern.

Third, you need solid data governance policies that reflect the act’s focus on fairness and non-discrimination. This means scrutinizing where your data comes from, ensuring its quality, and implementing technical measures to stop the AI from perpetuating or even amplifying existing societal biases. Employee training on ethical AI and the specific rules in PA 24-119 is also non-negotiable. And keep an eye out for more guidance from the Connecticut Department of Consumer Protection, you’ll need to stay on top of those updates for ongoing compliance.

For individuals in Connecticut who think an AI system has harmed them, these new rules offer a real path to compensation. Start by documenting everything. A denied loan, a job rejection, a questionable medical diagnosis, if you suspect an AI was involved, keep the records. The big advantage here is that the new transparency requirements may shift the burden of proof, making it easier to build a case. You should talk to a lawyer to see if PA 24-119 applies and if a claim can be built around a company’s failure to perform a proper impact assessment or mitigate known risks.

Impact on Injury Claims and Litigation

Public Act 24-119 is going to completely change how we handle AI-related injury claims in Connecticut. The old tools still work, we can still argue negligence, product liability, and breach of warranty, but now we have a powerful new weapon. We can file claims based on a company’s failure to conduct a proper impact assessment or to fix the risks they found, which gives us a much clearer shot at proving fault in these complex cases.

Imagine a hospital’s AI triage system misjudges a patient’s symptoms, causing a treatment delay that leads to a serious injury. Under PA 24-119, that hospital, as the AI deployer, could be liable for medical malpractice *and* for failing to properly vet its high-risk AI system. This change is huge for discovery. It means a plaintiff’s attorney can now subpoena the company’s internal impact assessments and risk mitigation strategies, documents that were previously almost impossible to get. The fact that companies are now required to create these assessments provides concrete evidence for litigation.

Because the act focuses so heavily on algorithmic bias, we’re going to see new kinds of lawsuits where injuries from discriminatory AI decisions can lead to claims that merge a classic civil rights violation with a specific violation of PA 24-119. This could open the door to stronger damage awards, including compensatory and possibly punitive damages, especially if a pattern of non-compliance or a willful disregard for the act’s rules is shown. The courts, particularly places like the Hartford Superior Court, are going to have to adapt to these new legal questions, and we expect the case law to evolve quickly as the first cases are tried.

We’re already gearing up for this new era. To win these cases, you’ll need a deep understanding of the AI technology itself, combined with a mastery of this new regulatory framework. Knowing the law isn’t enough anymore. The focus is shifting from proving general negligence to demonstrating specific failures in a company’s AI governance and risk management, as the new act outlines. It’s a nuanced but critical distinction.

And don’t forget the state itself. The State of Connecticut’s Department of Consumer Protection is the primary enforcement agency, with the authority to investigate violations and impose penalties like fines and injunctions. These regulatory actions can provide a strong foundation for an individual’s private civil claim, since a finding of a violation by the state would lend significant weight to a plaintiff’s case. The legal community needs to get ready, because the days of AI operating in a legal vacuum are over in Connecticut.

Connecticut’s Public Act 24-119 is a landmark piece of legislation. It creates new duties for developers and deployers and establishes new grounds for injury claims. For everyone involved in Connecticut’s AI world, from the companies building the tech to the people affected by it, understanding and adapting to this evolving legal field is essential for survival.