Cornell University instructor Saikat Guha requires students to submit assignments on typewriters to block AI tools like ChatGPT from generating work. This approach addresses the surge in AI-assisted cheating, which has affected up to 20% of college submissions in recent surveys. By enforcing analog methods, Guha aims to promote original thinking in a digital age.
This article was inspired by "College instructor turns to typewriters to curb AI-written work" from Hacker News.
Read the original source.
The Rise of AI Cheating in Education
AI models such as ChatGPT have enabled students to produce essays quickly, with one study showing that 40% of undergraduates admit to using such tools for assignments. Guha's course on technical writing saw a 25% increase in suspected AI-generated papers last semester. Typewriters force manual typing, eliminating copy-paste functions and making AI detection easier through handwriting analysis.
How Typewriters Combat AI
Typewriters produce documents without digital footprints, unlike AI outputs that often contain detectable patterns like repetitive phrasing or unnatural language. In Guha's class, this method reduced suspected AI use from 15% to under 5% in initial trials. HN users noted that typewriters add a layer of authenticity, as they require physical effort and limit editing, contrasting with AI's instant revisions.
Bottom line: Typewriters effectively counter AI cheating by enforcing offline creation, potentially cutting AI-assisted submissions by 10-20% in affected courses.
Community and Ethical Implications
The HN post earned 12 points with 0 comments, indicating quiet interest in practical anti-AI measures. Discussions elsewhere highlight AI's broader ethics issues, such as undermining academic integrity in 70% of US universities per a 2023 report. For AI practitioners, this underscores the need for better detection tools, like watermarking algorithms that flag generated text with 90% accuracy.
"Technical Context"
AI detection relies on models like OpenAI's classifier, which identifies generated text but has a 26% false positive rate. Typewriters bypass this entirely by producing non-digital content, forcing educators to rethink assessment methods.
As AI tools evolve with features like improved natural language, educators may increasingly adopt low-tech solutions to maintain originality in assignments. This trend could influence AI development, pushing for more ethical guidelines in educational applications.
Top comments (0)