EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Hacker News·Jul 27·5 min read
Professor's invisible prompt trap catches 32/35 students cheating with AI
TL;DR
A professor cleverly caught 32 out of 35 students cheating by embedding invisible instructions in an assignment prompt, leading to nonsensical AI-generated responses. This innovative approach highlights the challenges of academic integrity in the age of AI.
✦ Why It Matters
Educators should consider innovative methods to ensure academic integrity in assignments involving AI.
Key Takeaways
- Embed hidden prompts in assignments to test AI reliance.
- Monitor student submissions for nonsensical AI outputs.
- Encourage critical thinking by challenging students to verify AI responses.
- Explore alternative assessment methods beyond traditional exams.
How It Works
Dr. Gibson's method involved embedding invisible white text in the assignment prompt, instructing AI to include irrelevant content.
This clever tactic ensured that students who relied solely on AI would produce nonsensical answers, revealing their cheating.