web3 | crypto | regulatory law | policymaking | research

L

et's be upfront about AI: It’s so new right now that politicians and regulators are still trying to catch-up. The reality is that we need to distinguish between the regulators Congress thinks are in charge and the ones that

believe they are in charge.

This is an important distinction. Regulators get their authority from Congress. If Congress doesn’t grant regulators the authority and the budget, they aren’t permitted to regulate. But when innovation outpaces the legislative process, regulators may try to bandaid their existing authority to address emerging challenges. (Yes, it’s Congress’s job to pass laws, but they’re exceptionally slow at it). Unfortunately, these stopgap measures are often a poor fit because they’re not purpose-built.

Now let’s dive into the regulators and what they think:

Regulators that Congress has said are in charge of AI policy

1. OSTP (Office of Science and Technology Policy): While technically not a regulator, Congress has explicitly given OSTP the authority and earmarked budget to convene all the agencies and chair AI efforts. What does OSTP care about most? Its AI Bill of Rights. The AI Bill of Rights is directed more towards companies than to government use of AI. It outlines principles such as bias prevention, protection against unintended consequences, privacy protection, disapproval of user surveillance, and providing human alternatives to chatbots. Regulators will look to this document for how the White House wants them to think about AI.
   
2. Department of Commerce: Congress has appointed the Secretary of Commerce (a Cabinet-level position) to chair the “National Artificial Intelligence Advisory Committee.” It’s important to note that Congress chose Commerce over the other agencies such as the Department of Defense, Department of Energy, State Department, Department of Justice. This is Congress’ signal that when it comes to AI, it wants Commerce’s mission of economic growth and economic competitiveness to be the guiding principle.
   
3. NIST (National Institute of Standards and Technology): While technically within the Department of Commerce, Congress gave NIST the explicit task of developing the AI Risk Management Framework, (v.1 of which was released in January this year). Congress allocated real money for this work, starting with $64 million in 2021 and increasing to $94 million in 2025. What does NIST care about most? They want to create a voluntary, one-size-fits-all standard for companies to develop “trustworthy AI.” AI differs from traditional software, being inherently “socio-technical," a fancy word to mean that AI is influenced by social dynamics and human behavior. NIST believes companies’ use of its AI Risk Mitigation Framework will help implement proper controls and mitigate AI-perpetuated bias that “amplify, perpetuate, or exacerbate inequitable or undesirable outcomes.”
   
4. Department of Energy (DOE): Congress earmarked even more money to DOE to upgrade its infrastructure for AI and to award grants that support specific AI research, with a starting budget of $200 million in 2021 and growing to $262 million in 2025. Why? DOE oversees our nuclear infrastructure, our energy supplies, and our environmental program. AI could significantly enhance decision-making in these critical areas. And by awarding grants, the DOE will embed the NIST framework as conditions for any entity accepting those grants. Note how this incentivizes AI innovation with the measures the government wants in place, all without any regulations yet.
5. NSF (National Science Foundation): NSF is an agency known more for its role in funding research through grants than regulations. Congress allocated NSF an extensive AI grant budget, starting at $868 million in 2021 and rising to $1.05 billion in 2025. Huge. While not regulation, it’s worth noting the areas NSF prioritizes for AI research because these are the areas that Congress and the White House believe are most important for American competitiveness and growth:
1. Cybersecurity
   
2. Climate change
   
3. Understanding of the brain
   
4. Education (i.e., partially to address the concern that AI will kill jobs so we need to educate more people on how to use and develop AI)
   
5. Health
   
6. USPTO (United States Patent and Trademark Office): Interestingly, the US Patent and Trademark Office is the one federal agency that probably doesn’t need extra authority to regulate AI. Why? Because USPTO’s core mission is to grant patents and register trademarks. What does that have to do with AI? USPTO needs to address the crucial issue of patent ownership when the invention was accelerated by AI. The USPTO went out and sought public input on the issue, posted a request for comments from the public in the Federal Register, and they’ve received 69 comments so far.

Regulators that have self-appointed themselves to serve as the front line on AI

The following 4 regulators – CFPB, DOJ/Civil Rights, EEOC, and FTC – believe they are at the forefront of the government’s regulatory response to AI. However, neither Congress nor the White House have put them on the AI Task Force. So why do they think they’re in charge? They issued a joint statement to say that they are: “We also pledge to vigorously use our collective authorities to protect individuals’ rights regardless of whether legal violations occur through traditional means or advanced technologies.” Let’s get down to why these four regulators think they must crash the party:

1. CFPB (Consumer Financial Protection Bureau): The CFPB is known for protecting consumers in the financial marketplace (e.g., debt collection practices, credit card protection, payday lending oversight, student loan protections). So why does the CFPB have AI on its agenda? Because the CFPB is concerned that businesses are using or will use AI to make discriminatory and unlawful credit decisions. Moreover, the CFPB worries that businesses will absolve themselves by saying it wasn’t them, it was the tech.
2. Department of Justice (DOJ), Civil Rights Division: The Civil Rights Division of DOJ has a broad mandate to root out discrimination throughout our society, including social media platforms, banks, housing, and employment. They fear AI-driven business decisions tainted by built-in biases that lead to discriminatory outcomes. They want businesses to address those potentially discriminatory outcomes before they ever become a reality. As captured by the Assistant Attorney General who leads the division: “This is an all hands on deck moment and the Justice Department will continue to work with our government partners to investigate, challenge and combat discrimination based on automated systems.”
3. EEOC (Equal Employment Opportunity Commission): The EEOC is most concerned with how AI will change the landscape of employment decisions, in a discriminatory fashion. The Chair of the EEOC explicitly placed AI on its enforcement agenda, vowing “to use our enforcement authorities to ensure AI does not become a high-tech pathway to discrimination.”
4. FTC (Federal Trade Commission): Finally, we get to the FTC. If you’ve been following the story of FTC’s face-off with OpenAI, you might think that the FTC was in charge. And while the FTC may think so, too, Congress hasn’t been thrilled with the FTC’s enforcement agenda. So what does the regulator in charge of breaking up monopolies and protecting against fraud have to do with AI? The Chair of the FTC thinks that “AI tools can turbocharge fraud and automate discrimination,” and until she sees an “AI exemption to the laws on the books, the FTC will vigorously enforce the law to combat unfair or deceptive practices or unfair methods of competition.”

My closing thoughts: AI is still a bipartisan issue and politicians are trying in earnest to understand the tech without bringing threats or bans. Congress passed the National Artificial Intelligence Initiative Act of 2020, and they allocated a lot of money not to regulation, but to research, grants, and setting best practices. The White House also thinks this tech is important. Important enough for the President himself to fly to the geographic hub of the tech and convene a meeting about it. What other tech has enjoyed that special attention in your recent memory? My best piece of unsolicited advice is that if you’re working in AI right now, you can have an outsized impact if you meet the politicians where they’re at, using the vocabulary they understand.

Download the PDF here.

Connect

info@herminewong.xyz