AI and Easy Answers Threaten Genuine Learning, Expert Warns

The ease with which students can now use AI tools to complete assignments, such as writing lab reports, poses a significant threat to genuine learning, according to Professor Alon Korngreen, head of Bar-Ilan University's Brain Research Center. He argues that the "illusion of learning" created by AI-generated answers, which provide a sense of certainty without actual understanding, is more concerning than simple cheating.

Korngreen explains that learning fundamentally involves changing neural connections, a process that requires effort and, crucially, mistakes. When students copy from previous years or use AI, they bypass the struggle that solidifies knowledge. This "cognitive surrender" at the first sign of difficulty prevents the brain from engaging in the problem-solving that leads to lasting memory formation. He notes that students who rely on repeated reading or AI often feel more confident but perform worse on assessments than those who actively engage with the material.

He points to a 2026 MIT report that highlights these issues, warning against over-reliance on AI detection tools, which can be inaccurate and foster an arms race. Instead, the report suggests focusing on assessment methods that evaluate the learning process, such as oral defenses, cumulative portfolios, and assessing drafts, rather than just the final product. The MIT committee also advises against a complete return to traditional in-class exams, as they only measure what can be accomplished in a limited time, discouraging deep, long-term learning.

The professor emphasizes that AI can be a valuable learning tool, but only when used after the student has made a genuine effort. He advocates for a shift in how students interact with AI, moving from asking it to "do it for me" to asking it to "help me improve what I've done." This approach encourages the student to first generate their own work, then seek feedback to refine it, thereby fostering actual skill development. The decline in student attendance at office hours, classroom discussions, and study groups further exacerbates the problem, as these human interactions provide crucial "friction", challenges and differing perspectives, that AI often smooths over.