TL;DR

State-level AI legislation in the US is emerging as the most significant regulatory force, filling the vacuum left by federal inaction. These laws are more adaptable, targeted, and enforceable than any potential federal framework, setting de facto standards for the AI industry. While federal regulation stalls, states like California and New York are pioneering laws that will shape AI development and deployment for years to come.

The setup

Artificial intelligence is transforming industries, from healthcare to finance, but its rapid advancement has outpaced regulatory efforts. The US federal government has been slow to enact comprehensive AI legislation, with various bills stalled in Congress and a lack of consensus on key issues like data privacy, algorithmic bias, and accountability. Meanwhile, states are stepping into the breach, crafting their own laws to address the challenges and opportunities presented by AI. This patchwork of state regulations is not just a stopgap; it is becoming the primary driver of AI governance in the US.

The conventional wisdom is that federal regulation is necessary to provide a unified framework for AI, ensuring consistency and predictability across the country. However, the reality is that federal inaction has left a regulatory void that states are adeptly filling. This state-level activity is not merely a temporary workaround; it is a robust and dynamic system that is setting the pace for AI regulation.

The argument

1. State laws are more adaptable and responsive

State legislatures are often more agile than the federal government, allowing them to respond quickly to emerging AI technologies and issues. For example, California’s Consumer Privacy Act (CCPA) was enacted in 2018, setting a precedent for data privacy that many other states have followed. When it comes to AI, California’s Automated Decision Systems Accountability Act, introduced in 2023, aims to regulate AI systems that significantly impact individuals’ lives, such as those used in hiring, lending, and housing.

This adaptability is crucial because AI technologies evolve rapidly. Federal legislation, with its lengthy processes and bureaucratic hurdles, struggles to keep up. For instance, the Algorithmic Accountability Act, first introduced in 2019, has yet to pass, despite growing concerns about AI bias and transparency. In contrast, states like Vermont have already established an AI Task Force to study and recommend regulatory measures, demonstrating a proactive approach that federal regulators have yet to match.

2. Targeted and specific regulations

State AI laws are often more targeted, addressing specific industries or applications where AI has a significant impact. New York City’s AI bias law, which went into effect in 2023, is a prime example. It requires employers to conduct bias audits on AI tools used in hiring and promotions, ensuring that these systems do not perpetuate existing inequalities. This law is a direct response to the growing use of AI in HR and the documented biases in AI hiring tools.

Similarly, Illinois has enacted the Artificial Intelligence Video Interview Act, which mandates transparency in AI-driven video interviews. Employers must inform candidates if AI is used to analyze their videos and obtain consent. These targeted laws address specific concerns and provide clear guidelines for companies operating within the state, something a broad federal law might struggle to achieve.

3. Enforcement and accountability

State-level regulations are often more enforceable than federal ones, thanks to established state agencies and legal frameworks. For example, California’s Department of Financial Protection and Innovation has the authority to enforce the state’s consumer protection laws, including those related to AI. This means that companies violating AI regulations can be held accountable through fines, legal action, and other measures.

In contrast, federal enforcement of AI regulations is often hampered by jurisdictional issues and a lack of clear enforcement mechanisms. The Federal Trade Commission (FTC) has issued guidelines on AI transparency and fairness, but it lacks the resources and authority to enforce these guidelines effectively. State agencies, on the other hand, are better positioned to monitor and regulate AI within their jurisdictions.

4. Setting de facto national standards

While state laws are technically only applicable within their borders, they often set de facto national standards. Companies that operate across multiple states may choose to adopt the most stringent state regulations as their baseline to simplify compliance. For example, after California’s CCPA was enacted, several other states followed suit, and many companies adopted CCPA-compliant practices nationwide.

This trend is likely to continue with AI regulations. As more states enact AI laws, companies will face increasing pressure to comply with the most stringent requirements, effectively creating a national standard. This is particularly true for tech giants and multinational corporations that have the resources to adapt to varying state laws but prefer to maintain consistent practices across the country.

What the conventional view gets wrong

The conventional view that federal regulation is the only viable solution for AI governance overlooks several key points. First, it assumes that federal legislation can be enacted quickly and effectively, which is not the case. The current political climate and the complexity of AI issues make it unlikely that a comprehensive federal AI law will be passed anytime soon.

Second, the conventional view assumes that federal regulation would be superior to state laws, but this is not necessarily true. Federal laws are often broad and general, lacking the specificity needed to address the unique challenges posed by AI. State laws, on the other hand, can be tailored to local needs and concerns, providing more targeted and effective regulation.

Finally, the conventional view ignores the potential benefits of a patchwork of state laws. A diverse regulatory landscape can foster innovation and experimentation, as states try out different approaches to AI governance. This can lead to best practices that can be adopted by other states and eventually inform federal regulation.

The implications

For builders and users

For companies developing and deploying AI, state-level regulations mean a more complex compliance landscape. Companies must navigate a patchwork of laws, which can be challenging but also presents opportunities. By adopting a proactive compliance strategy, companies can gain a competitive edge by demonstrating their commitment to ethical AI practices.

For AI users, state laws provide greater protection against potential harms, such as bias, discrimination, and privacy violations. As more states enact AI regulations, users can expect more transparency and accountability from AI systems, leading to greater trust and adoption.

For regulators

State regulators are at the forefront of AI governance, setting the pace for the rest of the country. This presents both challenges and opportunities. On one hand, state regulators must grapple with complex and rapidly evolving AI technologies, often with limited resources. On the other hand, they have the flexibility to experiment with different regulatory approaches and set precedents that can influence federal policy.

For the industry

The current state of AI regulation presents both challenges and opportunities for the industry. While compliance with varying state laws can be burdensome, it also encourages innovation and the development of best practices. Companies that embrace ethical AI practices and proactively address regulatory requirements are likely to thrive in this environment.

What I’m watching next

1. The spread of AI-specific laws

In the next 6-12 months, I expect more states to enact AI-specific laws, particularly in areas like healthcare, finance, and transportation. These laws will likely focus on issues such as algorithmic bias, transparency, and accountability, setting new standards for AI governance.

2. Federal response to state laws

As state AI laws proliferate, the federal government may feel increased pressure to act. I am watching for signs of federal action, such as the introduction of new bills or the establishment of a federal AI regulatory body. However, given the current political climate, significant federal AI legislation is unlikely in the near future.

3. Industry response to state laws

Companies will need to adapt to the evolving regulatory landscape, and I am interested in how they will respond. Will they adopt a compliance-first approach, or will they push for federal regulation to create a more uniform regulatory environment? The industry’s response will be crucial in shaping the future of AI governance.

The bottom line

State-level AI laws are not just a temporary solution; they are the vanguard of AI regulation in the US. These laws are more adaptable, targeted, and enforceable than any potential federal framework, setting de facto standards for the industry. While federal regulation may eventually come, for now, states are leading the charge, shaping the future of AI governance. The implications are significant for builders, users, regulators, and the industry as a whole. As we navigate this complex landscape, it is crucial to recognize the pivotal role that state laws play in ensuring that AI is developed and deployed responsibly and ethically.