OPINION:
Thank you so much for your recent reporting (“‘It could kill us all by the end of the decade,’ an AI researcher says. He’s not alone,” (Web, Sept. 9) on one of the most important issues of our time.
Big Tech companies such as OpenAI and others are recklessly choosing to put human civilization in danger in their race to develop artificial intelligence general enough to replace humans in most jobs.
The fact that this could cause hundreds of millions of jobs to be lost is not the worst part, nor is the extreme cost in energy, land and water resources required to create all the needed data centers. Even the ability of AI to supercharge government and corporate surveillance and erode freedom isn’t the worst.
The worst is that they are trying to produce “recursive self-improvement” — AI agents that can create better AI agents. Once that happens, the machines are in charge.
A scientist once compared the Cold War arms race to two enemies locked in a gasoline-soaked room, both trying to hoard the most matches. Now there are more than two rivals and the matches are being designed to be able to light themselves.
There’s still hope for humanity to stay at the wheel of the car. Researchers and leaders have been quitting these companies for conscience. Lawmakers are considering legislation to ban the most dangerous kinds of AI. Safety legislation, research and international cooperation along the lines of President Reagan-era arms control treaties are vitally important.
As before, we must take action for the sake of future generations. If President Trump and Democratic and Republican leaders can work with corporate leaders and with other countries to step back from the brink, it would be one of the best things they could possibly do for their legacy, their descendants and ours.
JOHN DZIAK
Old Forge, Pennsylvania
Please read our comment policy before commenting.