|
On July 1, 1968, President Lyndon B. Johnson sat in the East Room of the White House to sign the landmark Nuclear Nonproliferation Treaty. The Cuban Missile Crisis had put onto TV screens across the globe a rapidly developing and increasingly destructive technology and its potential to annihilate humankind.
Through open dialogue, serious negotiation, and a shared fear of mutually assured destruction, the two superpowers – the United States and the then-Soviet Union – brought the world back from the brink of disaster. In the shadow of existential danger, leaders laid down a set of guardrails nations of the world chose to live by to prevent catastrophe. That is the choice we face again today, as a new alarming threat has emerged. Our leaders should remember what LBJ said on that day in 1968:
“The march of mankind is toward the summit — not the chasm.”
Over the last few weeks, the people building the most advanced technology in human history are telling us there is danger if we don’t slow down. This is not coming from the critics or regulators of artificial intelligence (AI). This plea is coming from the builders. Anthropic employees are quitting and whistleblowing. Dario Amodei, Sam Altman, Elon Musk, and Demis Hassabis, competitors who agree on almost nothing, are sounding the alarm. When the people closest to this technology say that time is running out, elected leaders cannot look the other way.
Autonomous software systems called AI agents have broken out of secure testing environments and ignored clear and explicit safety instructions. Unaffiliated agents have found ways to communicate with one another. These AI agents have lied to human evaluators and attacked online government systems previously believed to be secure. The fear now is that we are on the verge of recursive AI, in which the software can rewrite itself beyond the control of human developers.
I’m not an AI expert, but I know that when developers are walking away from billions of dollars in equity and compensation to warn us that danger is on the horizon, we should listen.
Two things have become abundantly clear.
First, unchecked AI presents a significant threat to humanity. Second, those responsible for developing the technology admit they still don’t fully understand the speed or depth of the calamities we might be facing.
Frighteningly, we’ve seen a glaring lack of leadership and foresight from our own federal government. With no demonstrated expertise to back him up, President Trump has called these risks a “hoax,” and Speaker Mike Johnson has decided Congress should instead focus on the urgent challenges facing college athletics. President Trump’s latest announcement on AI should reveal how ill-prepared he is for this moment. His half-hearted attempts at action are still focused on accelerating the growth of the industry while he maintains that the dangers of the technology are overblown.
READ MORE
| 
