Professor suspected AI-powered cheating on take-home midterms, makes finals in-person — only two students scored within 10% of their midterm score
A Brown University professor suspected that almost his entire class cheated on take-home mid-term exams using AI tools after they scored unusually high. In-person final exams showed that only two students scored within 10% of their midterm score, with just one getting a higher grade compared to their midterms.
The rapid proliferation and increasing sophistication of generative AI tools like ChatGPT are making it easier for students to leverage AI for academic work, forcing institutions to adapt. This event is a direct consequence of AI becoming readily accessible and powerful.
This incident highlights the immediate and widespread impact of AI on education and assessment, forcing a re-evaluation of traditional learning and evaluation methodologies. It also underscores the urgent need for new policies and tools to address AI-powered academic dishonesty.
Traditional take-home exams and assessments are increasingly vulnerable to AI-powered cheating, necessitating a shift towards proctored, in-person examinations or new assessment formats. The credibility of certain remote academic evaluations is diminished.
- · In-person education
- · AI detection software developers
- · Proctoring services
- · Traditional take-home assignments
- · Remote learning models without robust anti-cheating measures
- · Students relying solely on AI for understanding
Universities will accelerate the development and implementation of new academic integrity policies and AI detection tools.
There will be a push towards more project-based learning and critical thinking assignments that are harder for generative AI to complete convincingly.
The perceived value and rigor of online degrees and certifications may decrease unless robust anti-cheating measures are clearly demonstrated and enforced.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at Tom's Hardware