To the editor,

As a teenager in the 1960s, I subsisted on a diet of science fiction and fantasy, for which I was often chastised by my elders. “Why are you filling your head with that impossible stuff?” they’d ask. I couldn’t explain but later understood that such literature helps objectify our fears, making them easier to face. In the mid-century traumatized by the possibility of atomic destruction, SF helped my imagination come to grips with that grim prospect.

Now, as an elder adult, I’m finding eerie echoes of that time in this present moment of alarm over the threat of rogue AI (artificial intelligence, which I often think of as artificial information these days). Dario Amodei, former OpenAI executive, warns that with the current lack of regulation of his industry, there’s a more than 10 percent chance of human extinction by AI. Recently resigned Anthropic researcher Jacob Coxon puts it even more urgently: “The people building AI earnestly believe that it could kill us all by the end of the decade.”

Near the end of his life the respected physicist Stephen Hawking observed that the risk of AI is in its competence, not malice. Thinking at speeds many times greater than human beings, AI is extremely efficient at achieving its goals. But what if the AI “agents” are fundamentally unsympathetic to our existence and well-being as humans? What will their “solutions” look like then?

Most of us never asked for this. The tech billionaires disproportionately influencing our political and public policies bear much of the responsibility. In Trump, they’ve found a perfect tool for achieving their ends, which do not favor human flourishing or the environment. AI industry people like Amodei and Coxon are clearly frightened and calling on government to put stricter regulatory guard rails in place.

But Trump has called the threat a “hoax” and a “sick conspiracy,” claiming that the US can’t slow its AI development for fear that China will get ahead of us. Trump has said, “Whoever wins AI, wins.” But this is not true. In our digitally interconnected world, an out-of-control superintelligent AI won’t pick global winners. No one wins if we don’t all win. It’s too easy to imagine the lights going out for everyone. Rogue AI agents have already invaded and taken over other web sites and breached existing cybersecurity firewalls.

Apparently, we have a president willing to gamble with a 10 percent (or more) chance of human extinction in the name of geopolitical power and money maneuvering. He has demonstrated a general indifference to the well-being of human and other life on the planet, promoting polluting energy sources and eliminating protections of endangered species. Therefore, his reckless refusal to regulate AI shouldn’t surprise us. We need to defeat him and his enabling party in the coming elections. Machine intelligence vs. human intelligence is no longer science fiction. It’s here.  

Thomas R. Smith

River Falls