While AI assistants such as these comply with requests, they’re a long way from autonomous AI agents that weigh the pros and cons of multiple options and act without human oversight. Making that work in a campus environment is an immense challenge.
“Accuracy is always a problem if you’re going to fully automate things,” says Dan Harder, UT-Knoxville’s chief academic technology officer. Autonomous agents need human input to validate their assumptions and processes, he adds.
Harder notes that UT Knoxville’s technology teams work with the provost’s office to establish AI governance and guardrails. Policies define what kind of data AI tools can access and establish security and privacy controls to prevent, for instance, the use of private data in training AI models.
Education AI Agent Capabilities and Use Cases
Autonomous agents are still evolving, but the possibilities are growing.
Harder sees strong applications for using AI assistants in the campus HR department, including drafting job descriptions and summarizing the main points of extensive email threads. Just being able to use a tool such as ChatGPT to create the first draft of a report or memo can be a huge time-saver, he adds. AI assistants can help if your campus doesn’t have enough staff or budget to handle everyday challenges.
Similarly, researchers have noted a host of ways that agentic AI tools can potentially drive improvements in higher education. Agents will be able to gather data from multiple sources to assess a student’s progress across multiple courses. If the student starts falling behind, processes could kick in to help them catch up. Agents can relieve teachers and administrators from time-consuming chores such as grading multiple-choice tests and monitoring attendance.
The idea is catching on. Andrew Ng, co-founder of Coursera, launched a startup called Kira Learning to ease burdens on overworked teachers. “Kira’s AI tutor works alongside teachers as an intelligent co-educator, adapting in real-time to each student’s learning style and emotional state,” Andrea Pasinetti, Kira Learning’s CEO, says in an interview with The Observer.
Moreover, the Agent Laboratory at Johns Hopkins University crafted a framework for applying LLMs to review literature, write code and document research results. “Agent Laboratory significantly reduces research expenses, achieving an 84% decrease compared to previous autonomous research methods,” authors of an academic report about the lab write.
