AI cheating has already gone from asking a chatbot to write an essay to using tools that can make that AI-written work look human. We’ve previously talked about AI tools designed to help students cheat and the growing problem of AI detectors struggling to keep up, as well as how commonly students are already using AI for homework. Now the technology isn’t merely helping students do the work. In some cases, it can effectively be the student.
According to a New York Times report by Dana Goldstein and Alan Blinder, students can take things a step further and hand an entire online course over to an AI agent. As per the report, AI agents can be given simple instructions such as “log in and complete my quiz” and then operate inside online learning platforms. They can watch prerecorded lectures, take tests, write papers, and even participate in discussions with classmates.
That’s particularly problematic because online education is already built around digital, often asynchronous coursework. Students can complete lectures, assignments, and tests remotely, sometimes without ever interacting directly with their professors or classmates. More than half of American college students took at least one online class last year, according to the report.
And the AI doesn’t necessarily have to announce itself. The Times tested leading AI tools including ChatGPT, Gemini and Grammarly and found that none refused requests to write papers on a student’s behalf. AI companies also haven’t prevented agents from logging into platforms such as Canvas, Brightspace and Blackboard, although those services are researching ways to detect agent use.
The problem becomes even harder to police because these agents aren’t necessarily limited to one-off assignments. The online learning platforms themselves can look for clues such as how long a student spends on a task, but the NYT reports that there is currently no fail-safe way to block agentic cheating. Some educators have even asked AI companies to make their agents identify themselves when operating inside online courses, but those companies have resisted, citing privacy and security concerns.
That leaves universities with a particularly awkward problem. AI detection tools can make mistakes, making it difficult for professors to prove that a student cheated, while administrators may be reluctant to punish students without concrete evidence. The result is an inconsistent response, with some professors introducing in-person exams while others simply allow AI use.
Some educators are now redesigning online courses around work that’s harder for AI to fake, such as personal reflections and discussions about what students found confusing or interesting. Others are bringing certain exams back into physical classrooms.
But that creates a much bigger question about what online education actually means when an AI agent can attend the class instead of the person paying for it. The problem is no longer whether AI can help students cheat on an assignment. It’s whether a student can complete an entire degree without actually doing much of the learning. And for colleges that have spent years arguing that online education can make higher education more accessible, that’s a particularly uncomfortable problem to have.
Military working dogs charge into places the rest of us would flee. So, I understand why anyone would want to breed tougher, healthier dogs for the job.
Warhound, a new startup from Colossal Biosciences co-founder Ben Lamm and Anduril co-founder Palmer Luckey, has stepped out of stealth with $200 million to clone and gene-edit working dogs for American troops, police departments, and emergency services (via TechCrunch).
AI can describe a picture, identify objects, and answer questions about what a camera sees. But when a blind user needs a quick answer, getting an entire spoken essay about the scene isn't exactly helpful. Researchers at the University of Maryland (UMD) have found that AI-powered visual assistance can leave blind users doing too much work to get the information they actually need.
The study, titled Say It My Way: Exploring Control in Conversational Visual Question Answering with Blind Users, examined how people interact with Be My AI, an AI-powered feature of the Be My Eyes app. The researchers found that users often received lengthy descriptions that didn't adequately address their immediate goals, forcing them to ask follow-up questions or find ways to steer the AI towards a more useful answer.
Platforms tend to shrug off AI clutter until it threatens what they sell. For Etsy, that thing is trust in the maker.
To protect that, Etsy has banned AI-generated how-to guides and listings made entirely by bots. Most sellers must show at least one unretouched photo of the item (via Etsy Seller Handbook).