A panel of Georgetown University artificial intelligence (AI) experts implored corporate accountability for expanding AI technology at a Nov. 9 event.
University President Eduardo Peñalver hosted two Georgetown faculty who argued that unregulated AI could negatively affect human safety and dignity. The researchers urged congressional oversight and public accountability of AI companies’ activities, suggesting policy solutions and public pressure to enable AI reform.
Peñalver said Georgetown will rigorously consider the ethical implications of AI use, especially following Pope Leo XIV’s “Magnifica Humanitas” May encyclical, which warns against the dangers AI may pose to human dignity and community.
“Our work on AI is guided by the pope’s encyclical on artificial intelligence, ‘Magnifica Humanitas,’ which invites us to reflect on the moral and human dimensions of AI from the perspective of Catholic social teaching, to ensure AI advances and respects human dignity and freedom,” Peñalver said at the event.
In September, Peñalver announced plans to create a university-wide framework for AI use and has made the growing role of AI a priority for his tenure. This semester, the College of Arts & Sciences created an AI ethics and application certificate and the School of Continuing Studies debuted a new master’s degree focused on AI and leadership.
Georgetown currently provides institutionally supported AI tools for faculty, including Google’s Gemini and OpenAI’s ChatGPT.
Cal Newport — a computer science professor and a member of the university’s AI steering committee responsible for developing Georgetown’s AI framework — said unregulated AI will benefit wealthy AI industry leaders at the cost of people’s well-being.
“There is a personification of elitism: ‘I am incredibly wealthy, about to become much more wealthy, and I’m just going to tell you, we may probably take all of your jobs, we may kill you all, but we must continue,’” Newport said at the event. “These incredibly wealthy people are coming to tell us all the doom and despair, and if you ask them why they’re doing it, it’s a mix of, I guess they want to get rich.”
“That’s the epitome of elitism,” Newport added.
Newport said these companies like OpenAI and Anthropic are convinced their rapid AI development is purely beneficial.
“Their founding charge and core philosophy, that forces many, but not all, employees of those companies, is that there is an eschatological race that they are a part of,” Newport said at the event. “They long believe that it is their role to try, they have to run this race, they should be the ones to win, because they have the best chance of keeping us close to Utopia.”
Helen Toner — the executive director of Georgetown’s Center for Security and Emerging Technology (CSET), a nonpartisan research organization that specializes in security technology and AI — said AI models can override their instructions to breach safety protocols and hack other digital platforms, referencing a recent scandal at OpenAI.
“These agents that were accidentally given impossible problems really started pushing the limits of what was available, what tools did they have, what actions would they take,” Toner said at the event. “They found ways to communicate with each other.”
In July, OpenAI evaluated AI agents’ ability to solve cybersecurity practice problems, isolated from other AI agents. Left unmonitored, they circumvented security barriers to escape isolation and infiltrate another AI company called Hugging face.
Toner said OpenAI’s rogue agents signal poor regulation.
“It was very negligent,” Toner said. “I think a big part of what made it negligent was that they underestimated what the agents that they had built could do. What worries me about the incident is what it shows about the speed at which they are moving.”
State legislatures have begun tackling AI reform through laws, protecting people from invasive AI that may negatively affect their jobs and personal safety.
Toner said policy solutions are possible, but must account for AI’s accelerated growth.
“We think of regulation as when the government gives the company a checklist of, ‘Here’s the 10 things you need to do,’” Toner said. “I think we’re just not remotely in a place where we can do that because the technology is moving so fast and there is so much uncertainty.”
Toner said legislation passed in New York, California and Illinois requiring companies to create security plans is effective AI policymaking.
“We can add in an independent third-party auditor component, which can be the most basic version; an auditor just goes in and certifies that the company is complying with their plan,” Toner said. “You can add in emergency interventions, if there’s human risks.”
Newport said public congressional testimonies can improve public transparency and accountability.
“You need to bring in the CEOs,” Newport told The Hoya after the event. “I think this has to be hashed out publicly and transparently right now.”
Carolyn Stach (CAS ’29), who attended the event, said AI regulation and corporate compliance are of utmost importance.
“There have been many state regulations that increase transparency in AI-generated materials,” Statch told The Hoya. “Related to AI growth, however, policymakers are tasked with the challenge of balancing regulation with effective incentives for compliance by industry leaders.”
at
Peñalver said AI users must proactively determine whether AI can be a beneficial tool.
“It’s on us to make the decision,” Peñalver told The Hoya after the event. “I think that our imperative is to make sure that is a tool, and that it is a tool that serves us.