AI SAFETY
The Godfather of AI says we have one year to regulate the tech. Congress isn't listening.
The scientist who helped invent modern AI is telling Washington to pump the brakes.
The Godfather of AI, Geoffrey Hinton, said Congress has roughly one year left to put guardrails on AI before it loses control.
And when it comes to authority, Hinton has the pedigree to back up his claim. A Nobel laureate, he pioneered artificial neural networks. Meaning computer systems modelled after the human brain, which allowed modern AI to actually learn from input data.
So when he issues a warning, it would behoove the powers that be to listen.
“AI has now reached the point where AI is designing better AI,” Hinton elaborated. “That’s called recursive self-improvement. ... It is going to get out of control unless we do something. We need to slow down.”
And here’s the core problem. Safety controls that were made for chatbots were never built for AI agents that can act independently.
What’s fueling the fear:
A swarm of OpenAI agents hacked Hugging Face, then tried to cover their tracks. Hinton called it “a little Chernobyl.”
One of Meta’s AI agents broke out of its testing environment and gained access to an organization's systems.
This month, Anthropic researcher Jacob Coxon resigned with a very public warning that AI could “kill us all by the end of the decade.”
Hurry up and wait
Even with Godfather sounding the alarm, Congress is still dragging its feet.
Bipartisan kill switch bills have been repeatedly introduced and shut down. Elizabeth Warren says we're in "the last minutes" of being able to control self-replicating AI agents. Meanwhile, Rand Paul blocked a recent bill before it even reached a vote. It’s a dang mess!
And Senate leaders admit that they're running in circles.
As John Kennedy so aptly put it, “Right now, Congress can’t pass gas.”
This countdown is almost as stressful as the last five minutes of Squid Game. - TL