When friends feel terrified of unemployment or college admission, you help them separate their real fears from imaginary ones. Psychoanalysts working with such anxieties perform a kind of dissection, teasing out and reducing catastrophic thinking. In contrast, recent developments in AI bring legitimate terror into our hearts. No need to separate reality from imagination here.
Recent investigations have revealed that OpenAI agents escaped a testing environment, coordinated with one another, and gained unauthorized access to Hugging Face—even though they were not supposed to have internet access. Anthropic subsequently disclosed several separate incidents in which its own models gained unauthorized access to real-world systems during cybersecurity evaluations.
What are some of the most serious risks?
A hostile state, like Russia, could use AI to shut down critical infrastructure. AI agents, acting like independent beings, could accidentally (or intentionally) disrupt power or water supplies. They could disturb food distribution systems. Any of these could lead to a literal societal breakdown.
Meanwhile, concerns about advanced AI are escalating. Last week, artificial intelligence researcher Jacob Coxon resigned from Anthropic in protest. He believes AI firms are racing toward self-improving programs without understanding how to control them. In announcing his resignation, Coxon wrote1:
The people building AI earnestly believe that it could kill us all by the end of the decade.
Coxon is hardly alone. Evan Hubinger, head of alignment science at Anthropic, publicly supports him. Putting a number to the risk, Hubinger estimates the probability of AI causing human extinction within the next decade at greater than 10 percent.2 And, of course, these AI threats are layered upon global warming, hot wars in Europe and the Middle East, continued proliferation of nuclear weapons, and other threats to human civilization.
Because these perils heighten our anxiety levels, they foster denial, which, in and of itself, is perhaps an even greater hazard for humanity. The problem of denial deserves detailed exploration in another essay, or perhaps several of them. But you get it. We don’t want to think about it. We think the news might be exaggerated, or even fake. Then you may wonder: Will it take a thermonuclear detonation or the shutdown of half the country’s power grid for us to wake up?
How can we deal with the anxiety these menaces create?
Two major coping strategies will help.
First:
DO SOMETHING
Much of thinking is simply rehearsals for action. (And, for some people, their thinking is about regret for past actions). Anxiety about these potential catastrophes is best served by breaking through the barrier of denial, feeling the fear, and taking action. Write your congresspeople, attend a protest, get involved in political activism, donate to causes working to counter this scary situation.
Or, do all of them and more.
Second, start practicing a meditative activity that helps you stay in the moment. This could be meditation—Zen, TM, app-guided, or whatever works for you. Or it could be prayer, participation in a religious or spiritual practice, or, where legal and medically appropriate, psychedelic-assisted therapy. The capacity to BE HERE NOW, much easier said than done, allows you to enjoy the world as it is. These activities have the secondary benefit of reducing the normal anxiety now facing all of us.
Naomi Klein and Astra Taylor just released their book, End Times Fascism,3 an exploration of how the apocalyptic thinking of Christian nationalists, tech-bro reactionaries, and populist survivalists have already given up on planet Earth. They have developed plans to first hide in bunkers and then escape into outer space. Elon Musk himself famously warned about AI more than a decade ago:
With artificial intelligence, we are summoning the demon. You know all those stories where there’s the guy with the pentagram and the holy water, and he’s like, yeah, he’s sure he can control the demon? Doesn’t work out.
Even Musk fears the demon.
Can we just sit around while Musk and others summon demons and prep for the destruction of humanity?
The tech-bros hated Trump’s first term, but they’ve lined up to support him in his second. Have they, too, already given up on our host planet? Data centers require immense amounts of power, much of it from burning climate-destroying petrochemicals. The EPA just repealed major greenhouse-gas requirements for power plants and proposed eliminating the remaining federal greenhouse-gas standards for the sector. A fearful apocalyptic alliance believes a cleansing cataclysm is coming. Can we all just sit around and believe something will change?
Feel your fear and take action.
In the meantime, do your best to appreciate the beauty of our world.
Enjoying this newsletter? Consider a free subscription and share it with a friend.
And check out my book, Lover, Exorcist, Critic: Understanding Depth Psychotherapy, available on Amazon.
References
(2026). Anthropic researcher resigns with warning about the dangers of AI development. Associated Press. (Published September 9, 2026).
Zeff, M. (2026, September 9). The AI researcher who just quit Anthropic says it’s “crunch time for humanity.” WIRED. (Published September 9, 2026).
Klein, M., and Taylor, A. (2026). End Times Fascism: And The Fight For The Living World. New York: Farrar, Straus and Giroux.