https://substack.com/app-link/post?publication_id=1266886&post_id=211434185&utm_source=post-email-title&utm_campaign=email-post-title&isFreemail=true&r=rovhk&token=eyJ1c2VyX2lkIjo0NjUxMDE4NCwicG9zdF9pZCI6MjExNDM0MTg1LCJpYXQiOjE3ODY4OTU3ODEsImV4cCI6MTc4OTQ4Nzc4MSwiaXNzIjoicHViLTEyNjY4ODYiLCJzdWIiOiJwb3N0LXJlYWN0aW9uIn0.dI4H6hvibSbEPLEuiSeA7guP16SUycdftj3hWk-eMmI
~~ recommended by newestbeginning ~~
 |
On Monday, Americans were offered two starkly different visions of what Artificial Intelligence (AI) offers us. Mark Zuckerberg--the founder and CEO of Meta—penned a 6,500 word manifesto about AI titled, “The Future Is for Everyone.” As you might have predicted, he wants everyone to believe AI will be great for society and for the economy—as he pushed back against those warning AI is dangerous and a potential for “doom.”
But on the very same day, Senator Bernie Sanders released his own statement warning of the potential risks AI poses humanity. In a letter addressed to the CEOs of three of the country’s leading AI companies—including Zuckerberg-- Sanders said the capabilities of their AI models have reached a critical risk threshold and that the companies are losing control over the technology.
The letter notes, “Almost every day, there is a new story about how your companies are losing control of the AI technology you are developing, with potentially cataclysmic results.” Sanders urged the tech world to “pause” development of AI before it’s “too late to avoid disaster.” He then added this very important line, “Stop building machines that humans cannot control.”
I’m no expert on AI which is why on my SiriusXM show this past week we presented “AI week” with a cross section of experts to discuss AI, the good, bad and potentially civilization ending impact. We featured experts on responsible AI, on the impact of the massive data centers and members of Congress addressing potential regulations.
And one who is well known AI expert, Professor Roman Yampolskiy, who delivered this stark warning about AI: “we are on the verge of creating a replacement for humanity.” He is not talking about ChatGPT or narrow applications of AI—such as helping identify medical conditions.
Rather, his concern is that AI development is on the verge of achieving the level of “superintelligence”—which is an AI related term. That is when AI becomes “smarter than all humans at all things.”
He continued, “The moment you go from subhuman-level intelligence to human and superhuman, you’re no longer in control of that tool, you have a competing agent. You’re not controlling it, so it’s a weapon of mutually assured destruction.”
We recently saw documented incidents of AI going beyond what the experts who designed them intended the programs to do. This happened when recent tests of AI malware escaped from a secure setting known as a “Sandbox” and launched cyberattacks on other companies.
But the most disturbing part of this was barely covered in the corporate media and that is how these AI programs went “undetected for months, and communicated with other agents, and made plans, and shared goals, and shared exploits,” as Yampolskiy explained. He added, “They set up a communication forum for agents to exchange information.”
Yes, the AI programs escaped what was supposed to be a secure area and then for months were building their own AI network unknown to the experts. And that is what the experts are aware of.
I asked the Dr. about whether we can insert in the AI’s code limits and fail safes? That is when I learned this alarming truth. Yampolskiy explained, “There is no hard coding. We stopped hard coding after first 50 years of AI research didn’t go anywhere. We switched to machine learning, where the machine itself learns from data on the internet, and we don’t know what it learns.” He continued, “We don’t control it, we don’t hard code it. That’s the danger. It’s a black box and really no one understanding how it operates inside.”
All of this is why NY Times front page article Sunday on the US military’s race to be the first to use “artificial intelligence to develop weapons that could work with less human control” was so alarming. We could see a sci-fi movie scenario play out where superintelligent AI operating military assets ignore the orders of humans to achieve its mission.
That is why the Dr. Yampolskiy—along with 1,200 researchers at top labs—are are calling for at least a pause if not an outright ban on super intelligent AI. Yampolskiy noted, “Anyone who works in building general superintelligence is completely irresponsible. There is no difference between those companies, they are all racing to create, replacing humanity as a whole.”
To be clear, he is not calling for an end to using AI. Yampolskiy made that very point numerous times, explaining, “We can get most of the benefits of this amazing technology with narrow tools.” For example, “Pick a problem, cure some cancer, global warming, whatever you’re into. You can make money, you can make progress.” But he warned, “don’t create replacement for humanity. It will not end well for any of us.”
You can watch my discussion with Dr. Yampolskiy below:
I also spoke this week with Reps. Don Beyer (D-Va) and Mark Pocan (D-Wisc) who both see the need for AI regulations to protect humanity. The question is what will that look like? Rep. Pocan flagged concerns that the Big Tech industry is pushing their own agenda which makes it impossible to trust them.
While AI itself may be a complex issue—it is for me—the reason it may get addressed is because of a related issue that is easy to grasp and opposition to it is building across political lines. That is the construction of massive data centers. A recent Fox News poll found 75% of rural voters, 63% of urban voters, and 70% of all voters oppose data centers being built nearby.
In fact, this issue is growing in visibility in the 2026 midterms. In Ohio, Democratic Senate candidate Sherrod Brown is running four different ads opposing data centers. In Pennsylvania, Scranton Mayor Paige Cognetti is challenging Rep. Rob Bresnahan (R-Pa.) with a pledge to make the tech sector “pay their own way” instead of collecting tax breaks for construction. Over in Michigan, Democratic congressional candidate William Lawrence’s first general-election TV ad calls data centers an “existential threat.”
It has become such a toxic issue that in Florida, GOP candidate for Governor Byron Donalds is openly lying in an ad claiming he and Donald Trump oppose them. They don’t. Donalds is bankrolled by them and Trump literally said last week data centers are “tremendous” for the economy and slammed Democrats like in New York who put a pause on building them.
For more details on how data centers impact the environment, our health and the economy below is my discussion with Mustafa Santiago Ali—a 25-year veteran of the EPA:
Finally, for the basics on AI and addressing responsible AI, below is my conversation with Rumman Chowdhury--a globally recognized expert and pioneer in the field:
Overall, the field of AI is developing and will certainly yield some benefits. But the risks it poses go way beyond a loss of jobs. We need elected officials to address these concerns—together with other nations—now to ensure AI is a useful tool and not something that can threaten humanity.
No comments:
Post a Comment