TL;DR

State-level AI regulations in the US are emerging as the de facto standard-bearers for AI governance, filling a critical void left by federal inaction. These laws, while imperfect, are setting important precedents in privacy, transparency, and accountability. As states like California and New York enact their own AI policies, they are not only shaping the national conversation but also pushing federal legislators to act. However, the lack of federal oversight poses risks of a fragmented regulatory landscape, complicating compliance for businesses and potentially stifling innovation.

The setup

The rapid advancement of artificial intelligence (AI) technologies has outpaced the development of comprehensive regulatory frameworks. While the European Union has taken a lead with the AI Act, the US federal government has been slow to act, leaving a regulatory vacuum. This delay has significant implications for privacy, security, and ethical AI deployment.

Conventional wisdom suggests that AI regulation should be a federal matter, given the interstate nature of technology and the need for uniform standards. However, the current political gridlock and the complexity of the issue have stalled federal efforts. In contrast, state legislatures have been more agile, passing laws that address specific AI-related concerns. This state-level activity is not just a stopgap measure; it is shaping the future of AI governance in the US.

The argument

1. States are filling a critical regulatory void

The absence of federal AI legislation has created a regulatory void that states are increasingly stepping in to fill. This is not just a temporary workaround but a significant shift in how AI is being governed in the US.

For example, California’s Automated Decision Systems Accountability Act requires companies to assess the impact of their AI systems on consumers, particularly in areas like employment, housing, and finance. This law is one of the first of its kind in the US and sets a precedent for algorithmic accountability. Similarly, New York’s Automated Employment Decision Tool Law mandates bias audits for AI systems used in hiring and promotion, addressing concerns about algorithmic discrimination in the workplace.

These state laws are not just symbolic; they have real teeth. They impose fines and penalties for non-compliance, forcing companies to take AI governance seriously. By doing so, they are setting important precedents that could influence federal legislation and shape the national conversation on AI regulation.

2. State laws are more adaptable and responsive

One of the key advantages of state-level regulation is its ability to be more adaptable and responsive to local needs and emerging issues. AI technologies are evolving rapidly, and a one-size-fits-all federal approach may struggle to keep pace with new developments.

For instance, Illinois’ Artificial Intelligence Video Interview Act requires employers to notify candidates when AI is used in video interviews and to obtain consent. This law addresses a specific concern about AI’s impact on the job market and reflects the state’s priorities and values. Similarly, Vermont’s Data Broker Law requires data brokers to register with the state and disclose their data collection practices, reflecting growing concerns about data privacy and security.

These laws demonstrate how states can quickly respond to emerging issues and tailor their regulations to local concerns. This adaptability is crucial in the fast-evolving field of AI, where new technologies and applications are constantly emerging.

3. State laws are influencing federal policy

While federal legislation has been slow to materialize, state-level activity is increasingly influencing federal policy discussions. The success of state laws in addressing specific AI-related issues is putting pressure on federal legislators to act.

For example, the Algorithmic Accountability Act, introduced in Congress in 2021, draws on many of the principles and frameworks established by state laws. The bill requires companies to assess the impact of their AI systems and to take steps to mitigate any risks or biases. This is a direct response to the growing momentum behind state-level AI regulations and reflects the influence of state laws on federal policy.

Moreover, the Federal Trade Commission (FTC) has cited state laws in its guidance on AI regulation, indicating that state-level activity is having a ripple effect on federal agencies. In a blog post, the FTC noted that “states are leading the way in AI regulation,” suggesting that federal regulators are taking note of state-level developments.

4. The risks of a fragmented regulatory landscape

While state-level AI laws have many advantages, they also pose risks of a fragmented regulatory landscape. With different states enacting their own laws, companies may face a patchwork of conflicting and overlapping regulations, complicating compliance and increasing costs.

For example, California’s AI laws may differ significantly from those in Texas or Florida, forcing companies to navigate a complex web of regulations. This fragmentation could stifle innovation and create barriers to entry for smaller companies that lack the resources to comply with multiple state laws.

Moreover, a patchwork of state laws could create confusion for consumers, who may not understand their rights and protections under different jurisdictions. This could undermine trust in AI technologies and hinder their adoption.

Counter-argument: The need for federal preemption

One counter-argument to state-level AI regulation is the need for federal preemption. Critics argue that a patchwork of state laws could create regulatory uncertainty and hinder the development of AI technologies. They advocate for federal legislation that would establish uniform standards and preempt state laws.

However, this argument overlooks the fact that federal preemption could stifle innovation and prevent states from experimenting with new regulatory approaches. It could also lead to a lowest-common-denominator approach to AI regulation, as federal legislators may be reluctant to impose strict standards that could face opposition from powerful industry groups.

Counter-argument: The risk of regulatory capture

Another counter-argument is the risk of regulatory capture, where industry groups exert undue influence over the regulatory process. State-level regulations may be more susceptible to capture, as smaller, less well-funded state agencies may struggle to resist industry pressure.

However, this risk is not unique to state-level regulation and can be mitigated through strong transparency and accountability measures. Moreover, state-level regulations can serve as a check on federal capture, as states can enact stricter laws that reflect local priorities and values.

What the conventional view gets wrong

The conventional view is that AI regulation should be a federal matter, given the interstate nature of technology and the need for uniform standards. However, this view overlooks the advantages of state-level regulation and the limitations of federal action.

Federal regulators have been slow to act, and when they do, their efforts are often watered down by political compromise. In contrast, state legislatures can move more quickly and enact more ambitious regulations. Moreover, state-level regulations can serve as a testing ground for new approaches and ideas, providing valuable insights for federal policymakers.

The conventional view also assumes that a one-size-fits-all federal approach is the best way to regulate AI. However, AI technologies are complex and multifaceted, and a uniform approach may not be appropriate for all contexts and applications. State-level regulations can reflect local priorities and values, providing a more nuanced and tailored approach to AI governance.

The implications

For builders and users

State-level AI laws have significant implications for builders and users of AI technologies. Companies must navigate a complex and evolving regulatory landscape, which may require them to make changes to their AI systems and business practices.

For example, companies that operate in California must comply with the state’s AI laws, which could require them to conduct bias audits, obtain consent for AI-powered interviews, and assess the impact of their systems on consumers. Similarly, companies that operate in New York must comply with the state’s bias audit requirements for AI systems used in hiring and promotion.

These regulations could impose additional costs and burdens on companies, particularly smaller firms that may lack the resources to comply with multiple state laws. However, they also present opportunities for companies to differentiate themselves by demonstrating their commitment to ethical AI practices.

For regulators

State-level AI regulations are putting pressure on federal regulators to act. As more states enact their own laws, the need for federal oversight becomes increasingly apparent. However, federal regulators must be careful not to stifle innovation or impose overly burdensome regulations that could hinder the development of AI technologies.

Moreover, federal regulators should work closely with state agencies to ensure that state-level regulations are aligned with federal standards and priorities. This could involve establishing a framework for federal-state cooperation and coordination on AI regulation.

For the industry

The rise of state-level AI laws is a wake-up call for the AI industry. Companies must recognize that AI governance is no longer just a technical issue but a social and political one. They must engage with policymakers and stakeholders to shape the regulatory landscape and ensure that AI technologies are developed and deployed in a responsible and ethical manner.

Moreover, the industry should support efforts to develop voluntary standards and best practices for AI governance. This could help to build trust and confidence in AI technologies and provide a framework for responsible innovation.

What I’m watching next

1. The development of federal AI legislation

As state-level AI laws continue to proliferate, the pressure on federal legislators to act will only increase. It will be important to watch how federal lawmakers respond to this pressure and whether they can overcome political gridlock to enact comprehensive AI legislation.

2. The impact of state-level AI laws on innovation

It will be interesting to see how state-level AI laws affect innovation in the AI industry. Will these regulations stifle innovation or encourage companies to develop more ethical and responsible AI technologies? The answer to this question will have significant implications for the future of AI governance.

3. The role of international cooperation

As AI technologies become increasingly global, the need for international cooperation on AI regulation becomes more pressing. It will be important to watch how the US and other countries work together to develop common standards and frameworks for AI governance.

The bottom line

State-level AI laws are leading the charge in AI regulation, filling a critical void left by federal inaction. While these laws are not without their limitations, they are setting important precedents in privacy, transparency, and accountability. As states continue to enact their own AI policies, they are shaping the national conversation and pushing federal legislators to act. However, the lack of federal oversight poses risks of a fragmented regulatory landscape, complicating compliance for businesses and potentially stifling innovation. The rise of state-level AI laws is a wake-up call for the AI industry and a reminder that AI governance is a complex and multifaceted issue that requires a coordinated and collaborative approach.