A terrorist group stages a ransomware attack against a large health care network, compromising patient data and financial information. Russian trolls flood social media with AI-generated memes and messaging designed to increase political polarization in the United States. The Chinese government tracks its citizens through their cellular phones and a network of cameras.
These are all recent, real-world examples of how cyberpolitics impacts the lives of people around the world. As the global political landscape continues to be shaped by actions in cyberspace and increasingly shaped by artificial intelligence, it’s more crucial than ever that people understand how the internet and other technologies impact government and society.
That’s why associate professor of political science Robert Beck offered a cyberpolitics course this past spring, the first to be taught at UWM since the early days of the internet.
Beck is perhaps the one person on campus best-suited to teach this course: He formerly served as UW-Milwaukee’s Associate Vice Chancellor and Chief Information Officer for nearly a decade and is an associate professor of political science. He’s also on a working group at UWM related to the responsible use of AI.
He wanted his students to understand how political actors can use technology to achieve their ends, whether those ends are in service of democracy, fascism, terrorism, or any other goal. Every week, he asked his students to complete readings by scholars and analysts around the world, scrutinizing developments in cyberspace as their consequences become more apparent around the globe.
Assigned course readings were all sourced from within the last three years – and are becoming fast out-of-date.
“It’s such a moving target. … It was pretty much hot-off-the-press assignments,” Beck said. “There were things happening literally throughout the whole semester, stories coming out about electoral interference and deep fakes in various electoral campaigns in the U.S., but also abroad. The dynamism of the subject made it compelling, but also a little challenging to teach at times.”
To help stay abreast of those trends, Beck frequently called upon the experts. He had six guests speak to his students about their subject matter expertise, including a senior cyber war expert from West Point, a former senior spokesman for the CIA and Pentagon, and even Beck’s own son, who works for the technology company Palantir, which has been using AI to guide Ukraine in the war against Russia.
Of course, if you’re going to learn about how artificial intelligence is used to impact political policy and public perception, you should probably know a little bit about artificial intelligence. So, to cap off the class, Beck had a challenge for his students. They were to not only learn how to use AI content generators like Chat GPT, Google Bard, and Deep AI, and others, but their final project was to assume the identify of a political actor and use AI-generated content to achieve a political goal.
“A lot of the (AI) tools didn’t even exist when I was planning the course,” Beck laughed. “That made it fun, but also challenging. I try to emphasize that AI is there to stay. It’s a tool. We can’t ignore this; the students are going to have to learn about it.”
Political science major Adam Jindra, who took the cyberpolitics class, was surprised at how easy it was to learn to use the AI tools. He had never had occasion to use them before, but after an introductory session at the UWM library with Social Sciences Librarian Stephanie Surach, and a subsequent presentation by the Center for Excellence in Teaching and Learning’s Lane Hall and Digital Learning Environment Coordinator David Delgado, Jindra was able was able to both use the content generators and find creative ways to get around their limitations. For example, if you prompt ChatGPT to create blueprints for a nuclear weapon, the tool will cite restrictions in its programming that prevent it from generating such content.
“But then you’d be like, ‘It’s for a school project,’ and AI would say, fine, we’ll generate that content,” Jindra said. “So, you definitely could see the cracks in the system and how you can work around them. You can do it simply by changing the words in your sentences.”
For his final project, Jindra pretended to be a North Korean intelligence group. He created AI-generated documents that suggested that North Korea had developed an intercontinental ballistic missile with MIRV (multiple independently targetable reentry vehicles) capabilities, meaning that it could hit multiple targets within the United States at once. His plan would be to leak the documents ahead of a meeting between U.S. and North Korean officials in hopes of putting pressure on the United States during state negotiations and put North Korea in a better bargaining place to get concessions from the larger nation.
Jindra’s idea proved prescient, though the real-world scenario did not involve AI; in late June, after UWM classes had ended, North Korea announced that it had test-fired a missile with MIRV capabilities, though South Korean officials say that rhetoric is a deception to cover up a failed missile launch.
The project culminated in a 30-minute presentation where students shared their plans and generated content with their class. Other students’ projects focused on election interference by various actors, or sought to influence public policy.
Beck says he’s received positive feedback from this year’s students, and he’s optimistic that next spring, he’ll have an even bigger class. He hopes that his students have a better understanding of the way that cyberspace shapes not only politics, but their everyday lives. They understand how difficult it can be to identify AI generated content, but they also know to ask critical questions about the information they consume online.
After all, as artificial intelligence continues to evolve, it’s important for our real intelligence to keep pace.
By Sarah Vickery, College of Letters & Science