If it’s the End of the World, Think WOPR, Not Skynet.

If you’d prefer to listen instead of read, click on the video above or subscribe to our YouTube channel or podcast on YouTube Music, Apple Podcasts, or Spotify.

1. AI engineer Jacob Coxon resigned from Anthropic and wrote a viral X post that dominated the news cycle stating, “The people building AI earnestly believe that it could kill us all by the end of the decade.” Four days later, Dario Amodei, CEO of Anthropic, called for an AI development slowdown, a sentiment echoed by Sam Altman and Elon Musk with pushback from Jensen Huang and Mark Zuckerberg. Amodei acknowledged risks, but rejected so-called “doomerism.”

   • 🎙️ It’s hard to deny the real risks of frontier AI, but annihilation or civilizational collapse requires a certain level of human complicity or negligence. Maybe the risk is better understood with WarGames instead of Terminator. At this point it’s less like a maliciously sentient Skynet, and more like a misaligned WOPR. There’s a complexity to what’s happening that even the brightest minds can’t see clearly. I’m a little skeptical of voices with certainty, whether they call the warnings the “end of the world“ or a “hoax.” The fog of tech is like the fog of war, and the nature of both is unpredictability and uncertainty. The fog we see in real life limits sight. The fog of tech and war limits foresight. Which is why we believe transparency is the counter to the tech world’s illusion of intelligence. We not only need transparency for AI cyber incidents like the Hugging Face hack, but also transparency on training data, human reinforcement, recursive self-improvement, corporate financial or geopolitical incentives, and even anthropomorphic design. Calling AI “rogue” inadvertently implies it wasn’t programmed with goals. Saying AI “could kill us all” masks the liability or responsibility of those who develop it. Fear is too often used to secure positions of power. Unprecedented technology demands unprecedented transparency. The world doesn’t need superheroes or superintelligence. It needs ordinary men with self-control.

2. Two days after Dario Amodei’s call for a slowdown, Microsoft AI CEO Mustafa Suleyman released a “Humanist AI Code of Conduct” for AI models developed by Microsoft. The core premise of the code begins with the statement: “People matter more than AI.” Suleyman summarized Microsoft AI’s ten points on X that included: “AIs should not have rights or legal personhood,” “We're not racing to build a superintelligence that can slip its own leash,” and “Our AI should make you sharper, not dependent.”

   • 🎙️ The code of conduct was a refreshing reprieve from a wave of AI doomsday news coverage. Time will tell how this plays out in practice, but the code, at least in intent, aligns with our principles of Humanity, Privacy/Security, Integrity, Autonomy, and Transparency.

3. Last week the European Commission proposed the EU KIDS Act to ban social media and AI companions for children under 13, require parent-supervised accounts until they turn 15, and force platforms to make online minor accounts (social media, AI, and online games) “safe by design” by disabling features like “addictive design” and livestreaming. The EU would require users to use a state-run app to verify their age when registering on a new platform with an exemption for existing accounts where providers “can establish with a high degree of confidence that the applicable minimum age has already been reached.” The proposal would need approval by the European Parliament and EU member states to become law.

   • 🎙️ The EU KIDS Act is a significant leap beyond Australia’s social media ban with state-run age verification and fines for platforms that don’t meet the EU’s standards for minor account safety. If passed, the law could mark the end of an internet era where anyone could anonymously access platforms.

4. Apple has rolled out its newest device operating systems (OS version 27) now equipped with Siri AI, an AI assistant with a dedicated app. The latest versions represent Apple’s most significant Apple Intelligence integrations to date. It also includes “new and expanded child safety features” for iPhones like Ask to Browse, Communication Safety, and Time Allowances.

   • 🎙️ As with most new OS iterations, unless there are urgent security patches it’s usually prudent to wait to upgrade until the next incremental update (for example: iOS 27.1 or 27.2). The parental controls are welcomed additions, but their reliability remains to be seen.

5. Driver’s Training for the Online World is a video-driven conversation starter that helps parents prepare kids and teens for a lifetime of digital communication. Join us LIVE on YouTube on Thursday, October 22, at 7pm (CDT) as we guide you through Module 1 of our training in 60 minutes: the video, quiz, discussion questions, thought experiments, and action ideas.

   • 🎙️ Save the date, subscribe to our YouTube channel, share the details with friends, or test-drive the module for free on your own.