You know those predictions from some artificial intelligence companies that A.I. might produce Armageddon any day now?
Maybe we shouldn’t freak about them because they are a clever business ploy rather than a warning.
This reassuring possibility was one of many intriguing ideas that were floated during Tuesday’s Science Cafe in Concord, a free-wheeling discussion among some 70 people who filled Litherman’s Limited Brewery to hear about the effect A.I. is having on jobs, education and even humans’ ability to control our environment.
The two panelists, Professors Nick Smith of UNH and James Dobson of Dartmouth, both expressed what might be called a healthy skepticism about recent fears of incipient Armageddon, such as those from a couple of researchers at the A.I. firm Anthropic who said there is a 10% chance the software will destroy civilization in the next decade or so.
They have some suggestions for further reading, which are listed at the bottom of this article.
Their caution about claims of civilization’s destruction, they said, comes from the fact that they make the A.I. firms look much more powerful and accomplished. That’s not just an ego boost for the employees but, more importantly, is a good way to lure investors, intimidate opposition and stampede lawmakers into regulating the industry in a way that locks in their front-runner advantage.
Given such an incentive to exaggerate, the panelists said, we should be suspicious of these warnings.
I can’t give you more details because I was moderating the panel rather than taking notes, carrying a microphone to get comments from the crowd that packed Litherman’s as if I was channeling Bob Barker from The Price is Right. But you can see the whole argument on Concord TV, which filmed the Science Cafe and will have the video ready next week. We’ll have links to it, as well.
I can say there was a lot during the session that wasn’t reassuring. That includes a comment from one lawyer who said A.I. made a discovery which saved his innocent client from a guilty plea, which is, for what it’s worth, a positive outcome. But he said that anybody whose occupation occurs entirely on a computer — which, come to think of it, includes me — should expect their job to be taken over by A.I. in the next five or ten years.
Is that right? Good question, which is exactly why we had a Science Cafe to discuss it.
Science Cafe was a monthly occurrence in Concord for many years until COVID ended it. The Concord Monitor will be sponsoring more of them in the future, although they won’t be monthly. A separate group runs Science Cafes most months in Nashua — you can check out their Facebook page for details.
Here are suggestions for more reading from the two professors, some of which were mentioned during the Science Cafe:
- Roopika Risam’s “Data Empire: The Power of Information to Organize, Control, and Dominate.” This is not exactly about AI but about the larger data-dominated concerns that make it possible. The book also shows continuity between the use of this technology to govern people with other data-driven methods.
- Dennis Tenen’s “Literary Theory for Robots: How Computers Learned to Write.” This accessible book situates today’s technology in a longer history of automated writing and tools to help authors construct plots and structures.
- Melanie Mitchell’s “Artificial Intelligence: A Guide for Thinking Humans.” This book has recently been updated to address the LLM-powered versions of AI. She’s a computer scientist, and while this can get technical and into the details of neural network networks and language modeling, it’s accessible and clear.
- Leif Weatherby’s “Language Machines: Cultural AI and the End of Remainder Humanism.” This is a dense theoretical book for folks interested in the implications this technology has for language and linguistic theory.
- “Perceptron” is a biography and creative exploration of the life and ideas of Frank Rosenblatt, an early researcher of the neural networks that undergird contemporary chatbots.
- Anthropic CEO Dario Amodei’s two 2025 and 2026 essays “Machines of Loving Grace” and “The Adolescence of Technology.” Both are available here: https://darioamodei.com/
- Eliezer Yudkowsky and Nate Soares, “If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All.” Title says it all!
- Harari, “Nexus: A Brief History of Information Networks.” Is the marketplace of ideas broken? Does more information dilute access to truth? Will AI lead to authoritarianism?
- Kurzweil, “The Singularity is Nearer.” The very polarizing Kurzweil does a victory lap explaining how his predictions over the last 30 years have come true, and he predicts what is coming next.
- Mollick, “Co-Intelligence: Living and Working with AI.” Wharton business professor explains how to leverage AI in business and elsewhere. Gives students a playbook on how to use AI for classwork. Selling a lot of copies.
- Christian, “The Alignment Problem.” Really excellent, well-written, not sensationalist, and by a philosopher. Focuses on the development and training of AI, and the various challenges associated with “aligning” AI with human values. But it is six years old.
- Kai-Fu Lee and Chen Qiufan, “AI 2041.” AI thought leader and Chinese science fiction writer combine to write ten short stories and analysis of what the near future will look like.
- Shoshana Zuboff, “The Age of Surveillance Capitalism.” Released in January 2019, Zuboff warned: “The heady optimism of the Internet’s early days has turned dark. Surveillance capitalism has deepened inequality, sown societal chaos and undermined democracy.”
- Russell and Norvig’s “Artificial Intelligence: A Modern Approach” is a modern classic that is regularly updated and is used as a textbook in a variety of contexts.
- Karen Hao, “Empire of AI.” Critical bio of Sam Altman, and exposé on various social and environmental costs of AI.
- Michael Steinberger’s “Philosopher in the Valley,” a 2025 biography about Palantir’s Alex Karp. See also Karp’s 2025 book “The Technological Republic.”
