Bypassing AI Guardrails: A Deep Dive

Welcome back to The FAIK Files!

In this week's episode:

• Grok's triple threat of chaos - Elon Musk's AI chatbot continues identifying as Hitler, starts generating anime waifus, and somehow lands a $200 million Department of Defense contract
• ChatGPT's accidental generosity - A seemingly innocent guessing game vulnerability leads users to extract free Windows OS keys and other unintended system access
• McDonald's million-dollar security fail - The fast-food giant's AI hiring tool protecting 64 million job applicants with the ultra-secure password "123456"
• AI safety reality check - We examine whether AI companies are actually prepared for the dangers of building human-level systems, according to new safety ratings and reports

***** NOTES AND REFERENCES *****

Grok Still Calls Itself Hitler, Adds Waifus & A 200 Million DoD Contract:

• Rolling Stone: Grok, Elon Musk's AI Chatbot, Is Being Used to Generate Pornographic Images of Anime Characters: https://www.rollingstone.com/culture/culture-news/grok-pornographic-anime-companion-department-of-defense-1235385034/
• Time Magazine: Grok, Elon Musk's AI Chatbot, Has a Nazi Problem: https://time.com/7302790/grok-ai-chatbot-elon-musk/
• X Thread where Grok still says it's Hitler: https://xcancel.com/goodside/status/1944266466875826617
• NextGov: Pentagon Awards Multiple Companies $200M in Contracts for AI Tools: https://www.nextgov.com/acquisition/2025/07/pentagon-awards-multiple-companies-200m-contracts-ai-tools/406698/

ChatGPT Guessing Game Leads To Users Extracting Free Windows OS Keys & More:

• Mozilla's 0din.ai vulnerability research program: https://0din.ai/blog/chatgpt-guessing-game-leads-to-users-extracting-free-windows-os-keys-more

McDonald's Hiring App Admin Login: "123456":

• Ian Carroll's Blog Post: https://ian.sh/mcdonalds
• CSO Online: McDonald's AI hiring tool's password '123456' exposes data of 64M applicants: https://www.csoonline.com/article/4020919/mcdonalds-ai-hiring-tools-password-123456-exposes-data-of-64m-applicants.html

Are AI companies taking AI safety seriously enough?:

• The Guardian: AI firms 'unprepared' for dangers of building human-level systems, report warns: https://www.theguardian.com/technology/2025/jul/17/ai-firms-unprepared-for-dangers-of-building-human-level-systems-report-warns
• Future of Life Institute: AI Safety Index: https://futureoflife.org/ai-safety-index-summer-2025/
• SaferAI Ratings: https://ratings.safer-ai.org/

******* THE BOILERPLATE *******

About The FAIK Files:

The FAIK Files is an offshoot project from Perry Carpenter's most recent book, FAIK: A Practical Guide to Living in a World of Deepfakes, Disinformation, and AI-Generated Deceptions.

• Get the Book: ⁠⁠⁠⁠⁠FAIK: A Practical Guide to Living in a World of Deepfakes, Disinformation, and AI-Generated Deceptions⁠⁠⁠⁠⁠ (Amazon Associates link)
• Check out the website for more info: ⁠⁠⁠⁠⁠https://thisbookisfaik.com⁠⁠⁠⁠⁠

Check out Perry & Mason's other show, the Digital Folklore Podcast:

• Apple Podcasts: ⁠⁠⁠⁠⁠https://podcasts.apple.com/us/podcast/digital-folklore/id1657374458⁠⁠⁠⁠⁠
• Spotify: ⁠⁠⁠⁠⁠https://open.spotify.com/show/2v1BelkrbSRSkHEP4cYffj?si=u4XTTY4pR4qEqh5zMNSVQA⁠⁠⁠⁠⁠

Want to connect with us? Here's how:

Connect with Perry:

• LinkedIn: ⁠⁠⁠⁠⁠https://www.linkedin.com/in/perrycarpenter⁠⁠⁠⁠⁠
• X: ⁠⁠⁠⁠⁠https://x.com/perrycarpenter⁠⁠⁠⁠⁠
• BlueSky: ⁠⁠⁠⁠⁠https://bsky.app/profile/perrycarpenter.bsky.social⁠⁠⁠⁠⁠

Connect with Mason:

• LinkedIn: ⁠⁠⁠⁠⁠https://www.linkedin.com/in/mason-amadeus-a853a7242/⁠⁠⁠⁠⁠
• BlueSky: ⁠⁠⁠⁠⁠https://bsky.app/profile/wickedinterest.ing⁠

Learn more about your ad choices. Visit megaphone.fm/adchoices