AI Cheating Detection Industry Booms as Schools Struggle with ChatGPT

– ChatGPT, an AI chatbot, has gained popularity in the education sector, with reports of its use in college essays and high school projects.

– A survey conducted by revealed that 30% of college students admitted to using ChatGPT for written assignments.

– This widespread use of AI tools in education has created challenges for schools, educators, and students but has also led to the emergence of companies specializing in AI detection.

Companies like Winston AI, Content at Scale, and Turnitin offer subscription services that allow teachers to detect AI involvement in student work using a web dashboard

– and receive a probability score indicating the "human" or "AI" nature of the text.

– Most clients currently using AI detection services are individual teachers, but Winston AI is initiating discussions with school administrators at the district level.

Detecting AI-generated content often involves identifying specific features or "tells" that distinguish AI-authored text from human-authored text,

– such as overuse of certain words or phrases, lack of human flair, and the presence of typos.

Notable tells include "perplexity," which refers to the sophistication of language patterns, and "burstiness,"

– which refers to the repetition of words and phrases within a short span of time.

– The demand for AI detection services has surged since the release of ChatGPT, with Winston AI struggling to keep up.

– The company claims a 99.6% accuracy rate in detecting AI involvement.

– While some believe AI fears are overblown, industry experts like Annie Chechitelli of Turnitin argue that there will always be a tell, and new methods are being developed to unmask AI-generated content.