Johns Hopkins Economist Steve Hanke Says Spotting AI Cheating Is Easy — Here's Why

Steve Hanke, who has taught applied economics at Johns Hopkins University for nearly 60 years and served on President Reagan's Council of Economic Advisers, says AI cheating by students is a non-issue for experienced professors. Calling himself an 'old fox,' Hanke argues that detecting machine-written homework is straightforward, and attributes that ease largely to a paradox: students' own writing skills are so weak that AI-generated prose stands out immediately by comparison.

The core of Hanke's detection method is two-pronged. First, the quality gap — AI tends to produce cleaner, more structured prose than most students, including those at elite institutions, are capable of. Second, context — after decades in the classroom, Hanke has a granular sense of each student's grasp of economics, making it simple to flag submissions that don't match a student's demonstrated knowledge level. He acknowledges AI does require veterans like himself to stay vigilant, but frames it as a modest additional burden rather than a fundamental threat to academic integrity.

Drusilla Blackman, a former dean of admissions at both Harvard and Columbia and founder of Deans of Admissions — an advisory group helping families navigate elite university admissions in the US and UK — independently echoed Hanke's position. Blackman says it is nearly always evident when a student's submitted work fails to align with their established writing ability and critical thinking. The mismatch, she argues, is itself the tell.

Beyond relying on experienced judgment, educators across the board are adapting tactically. Teachers have reported developing AI-resistant assignments specifically designed to demand personal insight, lived experience, or real-time reasoning that chatbots can't convincingly fake. Some have gone further, reverting to handwritten in-class assignments entirely to remove the opportunity for AI assistance before it starts.