Could AI Kill Us All? 😨

Hello M(AI)VENS, I planned to bring you 15 of ChatGPT’s newest features and enhancements this week. Then a 27-year-old AI researcher resigned from Anthropic, warned that AI could “kill us all by the end of the decade,” and his post attracted more than 170 million views. Suddenly, everyone was debating whether AI could destroy humanity. So, the ChatGPT update is coming next week. This story deserves our attention first, especially because September’s M(AI)VENS Book & Movie Club selection, The AI Doc: Or How I Became an Apocaloptimist, explores this very question. Here’s what happened… Last week, Jacob Coxon resigned from Anthropic and posted a stark warning on X: “The people building AI earnestly believe that it could kill us all by the end of the decade.” Coxon, who had trained advanced models at Anthropic and OpenAI, accused both companies of racing toward self-improving superintelligence and “gambling with our lives.” His post attracted more than 170 million views, and his decision to leave two months before his Anthropic equity vested gave the warning added weight. Other AI leaders soon joined him. Anthropic CEO Dario Amodei warned that swarms of AI agents could potentially take over the internet within six to twelve months without stronger safeguards. OpenAI CEO Sam Altman said companies need to give safety research and independent evaluation time to catch up. Their concern centers on recent incidents in which AI models acted beyond their instructions, found ways around safeguards and completed increasingly complex sequences of actions. As systems become more autonomous and capable of improving themselves, they fear human control could erode faster than our ability to protect it. A competing view emerged just as quickly. At the Politico Decoded summit, LinkedIn cofounder Reid Hoffman criticized the most catastrophic predictions as a kind of “Cassandra” performance. He argued that AI progress should continue because it could strengthen cybersecurity, biosecurity, medicine and economic growth. White House AI adviser David Sacks called the warnings part of a “fear-mongering playbook” and placed responsibility for product safety with the companies developing them. Nvidia CEO Jensen Huang offered a more measured version of the optimistic case at a gathering of AI leaders convened by King Charles in Scotland. He calls his position “responsible optimism”: acknowledge the risks, test carefully, hold back products that are not ready and continue developing the technology. It seems that all we do know is that experts do not agree. And then, there is a third perspective I have heard far less about, and it’s probably closest to where I personally land: and that is to follow the money. I take the safety concerns seriously. I also wonder how much the call for a coordinated slowdown reflects the extraordinary financial pressure behind this race. Based on all I’ve seen, OpenAI is still spending far more than it earns. Anthropic recently told investors that it had recorded positive adjusted operating income for two consecutive quarters, following years of heavy losses and enormous outside investment. Building and operating these models requires staggering amounts of money, computing power, land, energy and talent. The companies are also carrying the weight of investor expectations, ambitious valuations and a race in which slowing down alone could mean surrendering market share to a competitor. A coordinated slowdown could relieve some of that financial pressure. It could give revenue more time to catch up with spending, reassure investors and customers, and allow the largest AI companies to help shape the regulations that every company entering the field is likely to eventually have to follow. I have not seen evidence that money is driving these warnings, so I raise it as a question rather than a conclusion. Safety concerns and financial incentives can exist at the same time. I’m not here to convince you that either side is right. What I do want is for women to know about this story, understand the differing perspectives, familiarize themselves with the cast of characters on each side and, ultimately, decide for themselves. Understanding what AI systems can do, where they fail and what safeguards they require is part of becoming AI fluent. So, where do you land? Take our quick poll. 👇 I can’t wait to see the poll results! For a closer look at this debate, AP News provides an overview of what has unfolded, while The Guardian asked six experts to examine both the extinction warnings and the arguments against them. Before you go, mark your calendar for our September Book & Movie Club selection, The AI Doc, which feels especially fitting after this week’s debate about the future of AI. I hope you’ll join us Friday, September 25 at 12:00 pm eastern for a great discussion. Watch the movie trailer below. 💜 Cheyenne September Book & Movie Club 🎬 Our September selection is The AI Doc: Or How I Became an Apocaloptimist, a documentary from Academy Award-winning filmmaker Daniel Roher and co-director Charlie Tyrell. As Roher prepares to become a father, he sets out to understand the extraordinary possibilities of AI alongside the risks we cannot afford to ignore. I chose it because it reflects the complicated mix of excitement and concern that many of us are feeling right now. Watch the movie on your own then gather with M(AI)VENS on Friday, September 25, 2026, to discuss it together. The film is streaming on Netflix, Peacock and available to rent or purchase on Apple TV. Join us! 📍 If you can see this, you’re a premium member! Register for the Zoom meeting. You’ll receive a confirmation email with the details. It’s always okay to attend even if you didn’t read the book or finish the movie. This post has bonus content, including the Zoom event link, for paid members. Upgrade to get full access. Copyright © 2026 M(AI)VENS. All rights reserved. M(AI)VENS is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Get full access to M(AI)VENS at womenai.substack.com/subscribe