UGA philosophers examine how people can preserve judgment, agency, and responsibility as AI increasingly shapes everyday decisions. 

This story is part of Humanities in Practice, a Franklin College of Arts and Sciences series exploring the value of the humanities in work and everyday life. 

Michael Zahorec and Nick Schuster, faculty members in UGA’s Department of Philosophy, study how AI affects human judgment. Together, they are exploring how people can preserve agency as AI increasingly influences practical and moral decisions. 

Generative AI responds to a request by producing content, leaving the user to decide what to do with it. Agentic AI completes larger tasks with limited—or sometimes no additional—human direction. Zahorec and Schuster focus on generative AI and recommendation algorithms, systems used by digital platforms that influence which possibilities people consider and which information receives their attention. 

Much of the cultural argument over the use of AI centers on academic cheating, disappearing jobs, and catastrophic risks. Schuster and Zahorec are studying a question already embedded in daily life: As persuasive machines offer to do more for us, what happens to human judgment? 

Schuster said he sometimes asks generative AI about his own research, but the responses are not always accurate. When the AI misattributes someone else’s work to him, the mistake provides a useful warning, he said. 

“Consider, then, whether you can trust the information it gives you about things you don’t know much about,” Schuster said. “Obviously no.” 

Their work suggests that AI may require a new kind of critical thinking. People still need to evaluate evidence and consider the source. They must now also understand how AI can shape the information they see, recognize the tendency to defer to that information, and accept responsibility for the conclusions they reach. 

Zahorec is a lecturer jointly appointed in the Franklin College of Arts and Sciences’ Department of Philosophy and Institute for Artificial Intelligence. His research focuses on AI interpretability and responsible use. Although generative AI can suggest possibilities, his “idea-generator” approach to using AI does not ask users to trust its output. 

Instead, Zahorec said users should treat AI output “merely as ideas,” leaving the users themselves responsible for verifying the information and separating useful suggestions from faulty ones. 

The approach applies well beyond the classroom, Zahorec said. A professional could ask AI to find weaknesses in a proposal, while a writer might use it to test a possible direction. When facing an everyday decision, someone could use AI to identify options they had not considered. Whether the response makes sense remains the user’s responsibility. 

Zahorec compares that process to writing philosophy. Building a sound argument requires a writer to anticipate criticism and decide which objections need an answer. Although some AI-generated objections may sound convincing, many will be weak. The writer still has to recognize the difference. 

“Exercising such virtues and skills is certainly the primary, if not the only, way to develop them,” Zahorec said.

Schuster studies the role attention plays in moral judgment. Recommendation algorithms already influence what people notice by selecting material for news feeds, search results, and entertainment platforms, he said. These systems help users navigate an unmanageable volume of information, but the algorithms’ choices about what to prioritize may pass largely unnoticed. 

“How judicious can we be about monitoring our information environment when algorithms deliver good (enough) content most of the time?” Schuster asked. 

Before users consciously consider an issue, a recommendation may already be shaping their opinions. If an algorithm repeatedly promotes inflammatory political content, for example, it can influence which people or problems seem threatening. 

Generative AI brings that influence into conversation. Schuster advises users to “play the role of Socrates” by questioning its claims, challenging its answers, and looking for inconsistencies. When people ask AI about subjects they know well, the system’s limitations often become clearer. 

This practice becomes especially important when AI participates in decisions affecting other people, Schuster said. His research on value alignment examines efforts to train systems to reflect human values, but those efforts face a persistent problem: Humans often disagree about morals. 

An AI system used in hiring, medicine, or public policy may make decisions efficiently and consistently, but its process still reflects values chosen by people. Those affected by AI’s decisions may have good reasons to question the results, Schuster said. Keeping humans involved preserves room to challenge those choices and take responsibility for their consequences. It also prevents technical performance from being mistaken for moral authority. 

Students could be well positioned to develop this broader form of critical thinking, Schuster said. AI gives them access to interpretations, objections, and questions that may once have been difficult to find. In the past, students might have accepted a claim without much scrutiny because it appeared in a textbook or came from someone with impressive credentials. AI gives them another reason to investigate it. 

Schuster uses class sessions to work on critical thinking and writing. Some exercises use pen and paper, while others incorporate generative AI. He believes younger students could also learn that the technology can be used well or poorly and that they retain “a lot of agency and responsibility.” 

Zahorec also sees education about AI’s influence as essential. He said AI companies can support research into explainability and design tools that encourage reflection, but a technical safeguard does not relieve users of the need for vigilance. 

That self-awareness may become a defining intellectual skill of this era, both researchers said. By remaining at the helm, students may become more deliberate about how they reach conclusions and more aware of their responsibility for what they choose to believe.