Trying to follow AI researcher Steve Omohundro as he described the best and perhaps only way to address the existential threat AI poses to our survival made my brain hurt.
But what he said — that we need to build the necessary guardrails into the technology itself — makes a lot of sense.
Meanwhile, The New York Times reports that “The newest and most powerful technologies — so-called reasoning systemsfrom companies like OpenAI, Google and the Chinese start-up DeepSeek — are generating more errors, not fewer. As their math skills have notably improved, their handle on facts has gotten shakier…
“Today’s A.I. bots are based on complex mathematical systems that learn their skills by analyzing enormous amounts of digital data. They do not — and cannot — decide what is true and what is false. Sometimes, they just make stuff up, a phenomenon some A.I. researchers call hallucinations. On one test, the hallucination rates of newer A.I. systems were as high as 79 percent.”
What is obvious is that AI literally has a mind of its own and that is what makes it so scary. Omohundro describes a scenario whereby an AI system might decide to invent, manufacture and distribute by drone a toxic virus that would be fatal to every human being on earth. Thus, his focus is on the guardrails we need to protect ourselves.
Merely comprehending AI, let alone the nature of its threat potential, can be a strain, but we really have no choice. That’s why I have been providing links every day not only to news stories but also to in-depth discussions.
To date I’ve listed around 1,000 references.
If you’re struggling to keep up, don’t feel bad. I’m struggling too. But it is vital that we try.
To contact your Congressional representatives about regulating AI, click on these links for your contact information in the House or the Senate.
(Thanks to John and Leslie.)
HEADLINES:
Hollywood is shaken by Trump’s tariff plan for the movie industry (CNN)
Here’s Where Trump Got His Idiotic Idea to Tariff Foreign Movies (TNR)
India Offers Zero-for-Zero Tariffs on Auto Parts, Steel From US (Bloomberg)
The U.S. Threat Looming Over Canada (Atlantic)
National Endowment for the Arts rescinds grants, dazing publishers and theaters (WP)
Trump administration offers $1,000 to undocumented immigrants to leave US (Guardian)
Cambridge's Harvard Square braces for Trump's war on higher ed (Axios)
Trump touts 'very strong' Alcatraz as Bureau of Prisons assessing reopening prison (ABC)
‘A slippery slope to eugenics’: advocates reject RFK Jr’s national autism database (Guardian)
Netanyahu vows to relocate Gaza’s population after security cabinet approves plan to ‘conquer’ the enclave (CNN)
India and Pakistan have significantly upgraded their military capabilities since the nuclear-armed neighbors clashed in 2019, posing increased risks of escalation even in a limited conflict, former military officers and experts say. (Reuters)
OpenAI reverses course, says its nonprofit will remain in control of its business operations (TechCrunch)
Zuckerberg’s new Meta AI app gets personal in a very creepy way (WP)
How China is still getting its hands on Nvidia’s gear (Economist)
Is It Too Late to Slow China’s AI Development? (FP)
A.I. Is Getting More Powerful, but Its Hallucinations Are Getting Worse (NYT)
Steve Omohundro on Provably Safe AGI (Future of Life Institute)
CDC Officials Announce Free Ice Cream For Everyone, Delicious Tasty Ice Cream, And Also There Is An Ebola Outbreak (The Onion)
No comments:
Post a Comment