E17: Hidden Prompts Hack: Trick LLMs for Contracts & Jobs (Top Unis Do It!)
The AI Cookbook: AI Tools | Enterprise AI | Leadership · Malcolm Werchota
Audio is streamed directly from the publisher (api.riverside.fm) as published in their RSS feed. Play Podcasts does not host this file. Rights-holders can request removal through the copyright & takedown page.
Show Notes
Universities like Columbia, Yale, and KAIST got caught hiding prompts in papers to force "positive review only" from LLM reviewers – a July 2025 scandal that's blowing up AI ethics. Malcolm deconstructs why it works (prompt injection hacks), shares his Bregenz family bot-trick stories, and drops a recipe to ethically test your own systems. Is this cheating or smart survival? Warning: This could change how you submit contracts or resumes forever. Beer recommended for the mind-bend.
Key topics:
- The hidden prompt scandal: Real examples from 17+ papers at top unis
- Deconstructing prompt injection: Why LLMs fall for white-text tricks
- Malcolm's kid-bot anecdotes: Family hacks that mirror uni tactics
- Ethical rants: Not illegal if unis do it – but what about business?
- Survival guide: Protect your contracts/jobs from AI reviewers