Strategic Ghosts, Synthetic Hosts
Welcome back to The FAIK Files!
In this week's episode:
- An Australian radio station created a fake Asian female host using AI
- The BBC resurrects Agatha Christie while a family member brings a murder victim to court
- We break down "Strategic Text Strings" - sequences of gibberish that can jailbreak AI systems
- AI recruitment tools might be making hiring worse, not better
Check out The Deception Project to learn about our upcoming Offensive Cyber Deception Masterclass and more.
Also check out Perry's new newsletter, Deceptive Minds: a newsletter about how we are fooled, how we fool ourselves, and what we can do about it. Subscribe on LinkedIn https://www.linkedin.com/build-relation/newsletter-follow?entityUrn=7319922626200510464
Want to leave us a voicemail? Here's the magic link to do just that: https://sayhi.chat/FAIK
You can also join our Discord server here: https://discord.gg/cU7wepaz
***** NOTES AND REFERENCES *****
Faking Diversity - AI Radio Host Controversy:
- Mediaweek article: https://www.mediaweek.com.au/backlash-over-arn-use-of-ai-radio-host-modelled-on-asian-woman/
- Original Blog Source: https://www.thecarpet.com.au/p/meet-thy-the-radio-host-i-dont-think
- Follow-Up with audio clips: https://www.thecarpet.com.au/p/radios-most-cynical-ai-experiment
- Mason's insights on radio industry cost-cutting and the vanishing of part-time jobs post-COVID
AI Resurrections:
- BBC resurrects Agatha Christie to teach writing: https://www.thetimes.com/culture/books/article/agatha-christie-ai-bbc-maestro-jeremy-vine-9t5nqvz69
- Family member resurrects murder victim to speak in court: https://www.404media.co/i-loved-that-ai-judge-moved-by-ai-generated-avatar-of-man-killed-in-road-rage-incident/
Strategic Text Strings & Jailbreaking AI:
- Previous episode on "AI SEO Optimization": https://www.youtube.com/watch?v=wm1yAfTjKzY
- Harvard Business School article: https://www.library.hbs.edu/working-knowledge/gen-ai-marketing-how-some-gibberish-code-can-give-products-an-edge
- Research paper: "Manipulating Large Language Models to Increase Product Visibility": https://arxiv.org/pdf/2404.07981
- Broken Hill: https://bishopfox.com/blog/brokenhill-attack-tool-largelanguagemodels-llm
- Broken Hill Github: https://github.com/BishopFox/BrokenHill
- Universal and Transferable Adversarial Attacks on Aligned Language Models: https://arxiv.org/pdf/2307.15043
AI Dumpster Fire of the Week - AI in Recruitment:
- 404 Media article on AI recruiter: https://www.404media.co/ai-recruiter-apriora-tiktok/
- Testing "AI Interviewers" with HeyGen (https://HeyGen.com/) and Tavus (https://www.tavus.io/)
Want to connect with us? Here's how:
Connect with Perry:
- Perry on LinkedIn: https://www.linkedin.com/in/perrycarpenter
- Perry on X: https://x.com/perrycarpenter
- Perry on BlueSky: https://bsky.app/profile/perrycarpenter.bsky.social
Connect with Mason:
- Mason on LinkedIn: https://www.linkedin.com/in/mason-amadeus-a853a7242/
- Mason on BlueSky: https://bsky.app/profile/wickedinterest.ing
