Image
School faculty share insights on how AI might influence policies, ethics, and governance
When it comes to artificial intelligence, what is ethical? How is this determined, and who regulates it?
This is exactly what School of Data Science faculty came together to discuss on September 13 at "The Future With AI: Policies, Ethics, and Governance." The event, hosted by UVA Lifetime Learning as a part of their More Than the Score: UVA Faculty Lecture Series, facilitated a conversation about the future of policy, ethics, and governance in the age of AI.
Renée Cummings, assistant professor of practice in data science, and Mona Sloane, assistant professor of data science and media studies, paneled the hybrid discussion on the evolving landscape of AI. Moderated by Michael Palmer, director of the Center for Teaching Excellence, the speakers shared their industry experience and insights on how to move forward with ethical AI regulation, development, and usage.
The panel began by discussing the issue that is perhaps closest to home: AI in education. Palmer highlighted the SAMR, or "Substitution, Augmentation, Modification, and Redefinition" model — a guideline for integrate technology into education — as a jumping-off point for how AI might aid educators. The panel saw some benefits in how AI might be utilized to personalize or tailor learning materials to individual needs. As Palmer explained, “Most pedagogical innovations, including technological innovations, only hit one of those levels, and mostly substitution or augmentation, whereas generative AI is a redefinition.”
Cummings and Sloane shared their experiences using AI in practical and innovative ways. Cummings described leveraging AI to turn pantry ingredients into creative recipes. “Sometimes when I'm home, and now with ChatGPT, I look at what I have in the pantry [and] it's like Iron Chef. What recipe could I come up with?” Sloane created a template that tracks her grant expenses and alerts her if something is overdue.
Beyond efficiency, both panelists emphasized AI’s role in learning and analysis. Sloane uses it to summarize student essays before providing her own detailed feedback: “I like to give extensive feedback to my students’ essays.... I ask ChatGPT to summarize this, and then I heavily edit that text.” But, Sloane argued that AI is more than just a tool. "It is a social infrastructure at this point," she said, noting that many societal decisions are becoming embedded in these technologies.
Cummings raised concerns about ethics and privacy. “AI continues to do things that we can call AI hallucinations," she said. "We're seeing things that are very, very questionable, but then we're seeing things that could be extremely spectacular. And that's the dichotomy, that's the double-edged sword that we've got to navigate.” She stressed the need for policy frameworks, noting that, in opposition to AI, government and governance are not real-time practices. "We've got to have that conversation about data. Data is an extension of our personhood," Cummings argued.
Panelists clarified common misconceptions about AI. Sloane distinguished between rule-based systems and modern deep learning, explaining that models can actually map relationships between data points in space and create predictions based on that.
Both emphasized AI’s limitations. “I do not think that we will end up in a situation where we have the LLM overlords, eradicating all of humanity," Sloane said. "We should not let [AI] take over other mega events and trends that we're seeing across society."
Cummings said that, for her, AI is the language of now. "There is nothing more extraordinary than human intelligence and creativity, and we will never get an AI to do that," she said.
For students navigating the rise of AI, Sloane advised recommitting to fundamental learning skills: “What I recommend my students do is to really, really recommit to learning how to learn," she said. "They have understood that this is their opportunity to learn how to think critically.”
In a rapid-fire session, panelists also reflected on AI’s potential: its superpower in geopolitics and forecasting, its positive disruption in healthcare, and its limitations in grasping an evolving context, which forces users to restart the chat and go back to ground zero.
In conclusion, Cummings highlighted the importance of ethical oversight in public sector AI. “We've always got to think about due process, equity, and fairness," she said. "Procurement is actually the place where we can enact some of these questions and where we can meaningfully think about stakeholder engagement and stakeholder expertise.”
The discussion underscored that AI’s integration into society is not merely technical — it is deeply human, ethical, and social. As these tools continue to evolve, the panel emphasized that innovation must be balanced with critical thinking, creativity, and careful governance.
In combination with the insights shared by the panelists, industry thought experts emphasize the importance of continuous dialogue and research in AI ethics and governance. Conversations about ethical AI are ongoing and should include everyone.