Stanford University ‘Boot Camp’ Teaches AI Congress

Artificial intelligence is perhaps one of the few topics capable of uniting Democrats and Republicans in Congress, at least when it comes to their shared desire to learn more about this rapidly evolving technology. Next week, congressional staffers from both sides of the House and Senate aisle will gather on the bucolic 8,000-acre California campus of Stanford University — not to admire the Romanesque architecture and campus neo-missionary, not to party, but to attend the Congressional Boot Camp on AI, hosted by the Stanford Institute for Human-Centered AI (HAI).

The three-day course will teach Congress attendees all about artificial intelligence – with sessions explaining what AI means for issues such as international security, the future of work, biases, privacy and health care. It includes field trips to Stanford labs for interactive experiments, as well as lectures by Stanford University professors and Silicon Valley executives. Participants receive a certificate of completion from Stanford University at the end of boot camp.

As Senate and House move on AI regulation, education is key

As the Senate and House race to catch up with the pace of AI development and tackle possible regulation, they have a lot to learn about these complex technologies, and their benefits and risks. This means teaching senators, representatives, and their staff everything from large language models (LLMs) and open-source AI to issues of AI safety, security, and ethics.

For example, Sen. Chuck Schumer (D-NY) is already planning a crash course in AI for senators this fall, which will include at least nine forums with top experts on copyright, labor issues work, national security, high-risk AI models, existential risks. , privacy, transparency and explainability, and elections and democracy.

According to Russell Wald, Stanford HAI’s Managing Director for Policy and Society, the boot camp is about providing attendees with the information they need to think critically about AI regulation and governance.

“For us, it’s not about educating people and telling them what to do,” Wald told Venturebeat in an interview. “It’s more about giving them the tools so they can have the critical thinking and analysis that would come from a lot of this.”

Stanford AI’s first training camp was held in 2019

Stanford HAI isn’t new to the concept of a “boot camp” to educate Congress — they held their first in 2014, which focused on cybersecurity. At the time, “there was just this lack of knowledge about cyber,” Wald said. “Someone on the intelligence committee said to one of my colleagues something like, ‘if you can take us on a scale of one to 10 from three to five in terms of general knowledge on this subject, you’ve made it service to the nation.”

This led to the first AI-focused boot camp in 2019. “There was someone from the ways and means committee who came to this program, now she’s the U.S. Trade Representative and cabinet member of Biden,” Wald said, presumably referring to Katherine. Tai.

These days, the AI ​​knowledge gap is still “pretty large” among congressional staffers, he said, but added that there is a great desire to learn more. “There are people who are informed, but I also find that they are in the middle of an evolving field,” he explains. “So for every minute you feel like you might have it, something changes – they know it’s important, they know they need to learn a lot, so there’s a level of humility to inside.”

This is important, he explained, because AI has now spread to more jurisdictions beyond congressional science or technology committees.

“[AI] will really affect people’s lives, it will affect health care, work, etc. “, did he declare. “So in that sense, it’s going to have to be people coming to the table – you don’t have to be a technologist, but that doesn’t absolve you of [being] a thoughtful and active participant [these issues].

A diverse cohort of Congress attendees learn about AI

Congressional staffers typically must apply for AI boot camp, which targets diversity by race, gender and party chamber.

“You don’t want everyone to come from the Senate Intelligence Committee, do you?” Wald said. The program also attracts those who work in a congressman’s home office, he explained: For example, Congressman Legislative Director Dan Crenshaw (R-TX) attended last year, as well as staff members of Sen. Cory Booker (D-NJ) and Sen. Padilla. (D-CA) and the executive director of the Congressional Black Caucus.

It is also important that the group be bipartisan, he added. “It’s something I’m personally looking for – it’s such powerful technology that if we let the vitriolic politics that we seem to be surrounded by lead the way, it’s going to be a huge problem,” he said. “I’m not of the mindset of we teaching Democrats, or we teaching Republicans. We teach those who want to learn and help them understand.

This cohort of 2022 heard from people like Fei-Fei Li, professor of computer science at Stanford University and co-director of Stanford HAI; Peter Norvig, Distinguished Education Scholar at Stanford HAI and Research Director at Google; and Percy Liang, associate professor of computer science at Stanford and director of the Stanford Center for Research on Foundation Models.

A multidisciplinary approach to teaching AI at Stanford

Wald said Stanford AI experts work hard to meet decision makers where they are, especially when they have so many issues on their plates. “It’s about, what do we do to effectively reach them so that they’re really informed by this?”

That’s one of the reasons, he added, that Stanford isn’t doing the AI ​​boot camp in Washington, DC.

“I won’t do it in DC because life is going to kick in,” Wald explained. “All of a sudden it’s like ‘my kid has a cold today’ – there’s a huge difference in commitment when you say you’re getting on a plane and you’re going to come here.”

The program begins with a basic tutorial on what AI technology is. “We’re not telling people to pull out their laptops, we’re going to start building a model,” Wald said. “These types of programs are actually counterproductive. We teach them to understand what technology is, what it can and can’t do, and its impact on society – that’s what we really want them to get out of it.

Wald admits that Senate and House staffers also have access to AI information from Big Tech companies such as Microsoft and Google. But, he said, there are important differences.

“First, we are an academic institution, so we teach people,” he said. “But also, Microsoft is ultimately going to be a product lab – their goal is to have commercially viable products that will benefit a market. We have many teachers who have unique and specific experience in a field, so [what we offer is] multidisciplinary, looking at things from a very broad point of view.