Artificial intelligence (AI) is no longer just a tool for convenience – it is making decisions that impact people’s lives in hiring, housing, healthcare, credit, education, and even the criminal justice system. The way we regulate AI today will define whether it serves as a tool for progress or becomes a force for discrimination, manipulation, and harm. That’s why I strongly support Senate Bill 2, An Act Concerning Artificial Intelligence, which takes an essential step toward ensuring fairness, accountability, and transparency in AI systems used in Connecticut.
Accountability in AI Decision-Making
One of the most pressing concerns is that AI is often used to make decisions with no clear line of responsibility. If an AI system denies someone a mortgage, rejects their job application, or determines their risk level in the criminal justice system, there must be a person who can be held accountable for that decision. AI cannot be a faceless authority making life-altering choices without oversight.
Under Senate Bill 2, organizations using AI to make significant decisions about individuals must have human accountability built into their systems. This means that a real person – one who can be held responsible – must oversee AI-driven decisions and ensure they are fair, unbiased, and legally compliant. Furthermore, that person must have a legal obligation not to discriminate against individuals based on race, gender, disability, or any other protected category. AI should never be a shield for discrimination – it must be designed to prevent it.
Transparency: People Deserve to Know
If a company, a government agency, or any organization is using AI to make decisions that impact a person’s life, that person should be informed before the system is adopted, while it is in use, and when they have concerns about bias or discrimination.
People deserve the right to know when an AI system is being used to determine their job application, loan approval, college acceptance, or even how they interact with government services. They should also be given clear steps to appeal a decision if they believe the AI has wrongly or unfairly judged them.
Preventing Harmful AI-Generated Content
Another major concern is the rise of synthetic images, deepfake videos, and AI-generated content. While AI can be used for creative and harmless purposes, it is increasingly being weaponized to spread misinformation, harass individuals, and manipulate public perception.
For example, AI-generated images have been used to create harmful fake content, from deepfake revenge pornography to fabricated political attacks meant to mislead the public.
This is not just unethical – it is dangerous. Senate Bill 2 requires clear labeling when AI-generated images or videos are used so that people can distinguish between real content and manipulated media. Additionally, using AI to maliciously create fake images or videos of someone with the intent to harm them should be illegal. No one should have their image stolen, altered, and weaponized against them.
A Path Forward
Senate Bill 2 is not about stifling innovation—it is about ensuring that AI serves the public good rather than becoming a tool for exploitation. Connecticut has the opportunity to lead in creating AI laws that are strong, commonsense, and fair. We must act now to ensure accountability in AI decision-making, transparency in AI usage, and protections against harmful AI-generated content.
The future of AI should not be decided by tech companies alone—it should be shaped by the people it affects the most. That is why I am proud to support this legislation, and I urge my colleagues to join me in building an AI future that is ethical, fair, and accountable.
Summary
- AI is increasingly making critical decisions in various sectors, necessitating regulation.
- Senate Bill 2 emphasizes human accountability in AI decision-making.
- Transparency is crucial for individuals affected by AI-driven decisions.
- The bill addresses the dangers of AI-generated content and misinformation.
- Connecticut aims to lead in fair and commonsense AI legislation.