Tell us a little about your career journey: How did you come to work in AI policy?

I first interacted with AI policy during my PhD in artificial intelligence at Carnegie Mellon University. I encountered the topic through student government and science policy organizations. I took a great interest in it, and pursued every opportunity to get involved with policy issues. For example, when I was a graduate student there were debates about the visa status of STEM students and I got involved in advocacy campaigns. As a result, I got to help on a couple of Congressional bills, including legislation related to tax rates and jobs. Eventually, I ended up working on AI-related bills as well. Through that work, I crossed paths with the Federation of American Scientists, which is how I ended up with my current job.

What are some of the current AI policy challenges you’re working on?

I think there are a lot of technical challenges that people need to figure out. We keep talking a lot about standards for AI systems, but defining standards is very hard. Cybersecurity for AI systems is another crucial challenge. There’s not enough technical talent supporting the government and the right organizations. 

How do you make sure that the benefits of the technology are widely shared is also an important question. We do not want to end up in a similar position to where we are with broadband and digital connectivity. The prosperity that those technologies brought was restricted to certain regions while others were left behind. We have to make sure that we do not find ourselves in the same place with regards to AI three decades from now.

Another thing we need to think about is AI regulation that guards against current harms as well as potential future emerging threats. One analogy is that preparing for lower grade earthquakes can help us prepare for larger ones. We’ve never had a category six hurricane, but it could happen. We can prepare our infrastructure now and incentivize innovations now to protect against it. For AI, we similarly need to think about what preparations we can make right now.

The government could create AI foundation models for its own use. Attracting top talent to the US is another obvious place to start encouraging innovation.

What advice do you have for those interested in a similar career path?

No opportunity is too small, and no impact is too small. Don’t think that you’re going to fix everything tomorrow, because that’s a recipe for burnout. Think of what incremental changes you could make today, right now. Sometimes changing a single word in a sentence makes a lot of difference. Start from what you know, and think: what’s the smallest thing you could do? Everybody wants to hit home runs, and sometimes you just need to get a steal.

What skills do you think are important for success in AI policy, and how could readers acquire them?

Entrepreneurship is important. You have to be relentless, you have to go out and learn. You can’t just sit around and hope that things will come to you. That’s true for policy in general. You have to have a willingness to learn and a willingness to recognize that your prior beliefs might be wrong. It’s important to recognize false beliefs quickly. Congress is not such a fast-paced environment, but being fast is still important in testing out and getting better at entrepreneurial skills. 

Are there any programs, resources, or books you’d especially recommend for those interested in AI policy?

• The Structure of Scientific Revolutions by Thomas Kuhn
• Robotica: Speech Rights and Artificial Intelligence
• Chapter 20 of The Armed Forces Officer (to improve at writing)