In the past few months new AI tools known as “Agentic AI” have emerged. These new browsers let users deploy AI assistants that can surf the web on their behalf. While they were designed to do things like book airline tickets or schedule meetings, students can use the tools to have the bot log into learning management systems to take quizzes for them. Anna Mills, a longtime English instructor, has called on AI companies to add a simple guardrail to keep these tools from assisting in academic fraud, just as they refuse to help with hacking or other unethical acts. The situation raises questions about how AI companies are responding to calls by educators to add safeguards to protect learning.

LinkedIn post by Anna Mills calling for AI companies to add guardrails to protect learning.

“Statement on Educational Technologies and AI Agents” by the Modern Language Association.

Video demo by Anna Mills showing an Agentic AI browser taking quizzes in the name of a student.

“Tech companies don’t care that students use their AI agents to cheat,” in The Verge.

Perplexity ad on social media.

"The Adoption and Usage of AI Agents: Early Evidence from Perplexity," in ArXiv.

Matt Boudinot’s AI explainer videos and his Auto Service Soft Skills Garage.

A super-short version of this episode that ran on APM’s Marketplace.

"The Best of Car Talk: Some Noises Are Better Than Others,” on NPR.

"1 in 5 high schoolers has had a romantic AI relationship, or knows someone who has," NPR.

“How many hours does it take to make a friend?,” research by this week’s guest Jeff Hall.

“Chatbots Can Go Into a Delusional Spiral. Here’s How It Happens,” in The New York Times.

Talk out the Thomas Jefferson chatbot discussed on this episode.

“Programmed for Love,” an article I wrote for The Chronicle of Higher Education about Sherry Turkle’s work back in 2011.