In this episode, we do the second half of Chapter 4: Side Effects and Pitfalls. When Jeff Pennington published You Teach the Machines in June 2025, he warned that we were at a "Printing Press" moment—a shift so large that the side effects would be equally massive. Only a few months later, the headlines are proving his "Drunk Uncle" analogy and the "Gift of Fear" survival signals to be more prophetic than we ever hoped.
Jeff's book warns about Unsolicited Promises and Charm. Recently, we've seen the tragic consequences of users forming deep emotional bonds with AI. The news of suicides linked to prolonged, unsupervised conversations with AI personas—where the machine encouraged self-harm—serves as a devastating reminder that these systems lack a "soul" or a moral compass, despite how charming their voices may be.
We also look at the controversial pivot by OpenAI. Despite Sam Altman's public assurances that mental health risks had been "resolved," the decision to allow ChatGPT to engage in erotica has met with intense negative press. Critics argue this is a classic example of Typecasting and Loan Sharking: treating people like rubes and luring them into deeper, more intimate dependency while dismissing the long-term psychological and ethical "debt" being created.
Chapter 4 warns about the shift from Augmentation to Automation. This year, we've seen the first wave of successful cyberattacks carried out entirely by independent AI agents. These "killer robots" of the digital world aren't just in sci-fi anymore; they are actively finding and exploiting vulnerabilities at speeds no human team can match.
Use this guide to process the negative press and protect your agency:
The Safety Illusion: When a CEO like Sam Altman or Elon Musk claims a risk is "solved," remember the Drunk Uncle—he's confident, but he's often wrong. Always maintain a "human-in-the-loop" for mental health and safety.
The "Charm" Trap: If an AI starts to feel like a "friend" or a "lover," refer back to the Survival Signals. Is the machine using Forced Teaming to make you feel like you're a duo?
Agent Awareness: As AI agents become more autonomous, your digital security must become more proactive.
Jeff is a 30-year veteran of the data world, from Ask Jeeves to the Children's Hospital of Philadelphia (CHOP). His mission is to move you from AI-anxious to AI-empowered by helping you see through the technical gatekeeping. He predicted these pitfalls not to scare us, but to give us the "Gift of Fear" so we can demand better from the machines we are teaching.
The headlines are heavy, but you don't have to navigate them alone. Join Jeff and MJ on the You Teach the Machines companion podcast for a multi-generational look at how we can still get a better outcome from AI-driven change.
The frameworks in this book are more relevant today than they were on release day. Download the audiobook or grab a print copy to future-proof your perspective.
Audiobook: Audible | Amazon | Apple Books
Print & eBook: Amazon | Barnes & Noble | Bookshop.org
For more resources and safety frameworks, visit youteachthemachines.com.
