Red Alert
By
Leonard Zwelling
https://www.nytimes.com/2026/09/12/technology/anthropic-dario-amodei-ai-slowdown.html
https://www.nytimes.com/2026/09/11/us/politics/cia-sept-11-intelligence.html?searchResultPosition=1
All of these articles were on The New York Times website on September 12.
The first describes how the AI company Anthropic was able to block attempts to use its large language model Claude to create a biologic weapon. That anyone was using AI for this purpose is alarming.
The second describes how the CEO of Anthropic, Dario Amodei, is warning that the rapid development of the power of these AI-driven models ought to be slowed down to make sure none becomes a threat to humanity. There have been several AI experts who have quit their jobs in protest because the guardrails on such potentially harmful models are inadequate. One of these former employees has assessed the chance of AI killing us all at 10% within a decade. That’s pretty high. This too is alarming.
The final story is about the newly released CIA documents that tried to warn Presidents Clinton and Bush about the threat posed by Al Qaeda to the United States as early as 1998 including the potential use of explosive-laden airplanes to harm Americans. I guess this should have been alarming, but wasn’t alarming enough.
If all of this does not send a chill down your spine, then you are not paying attention. The warnings were there before 9/11. Our leaders lacked vigilance and a belief that our adversaries would have vivid imaginations to do us great harm. America was not on guard. Our leaders were inadequately alarmed.
Previously, this blog has urged those in charge of the major AI firms, representatives of the government, and any other expertise that America can bring to bear on the threat AI poses, meet immediately to establish policies and processes to assure the American people that AI will be developed for the good of mankind by establishing ways to prevent its use by those who would harm us. Yes, it may be true that our enemies like China would use AI against us. Even if the U.S. creates rules, others outside of our purview need not follow our example. How to prevent that and protect the United States needs to be a major focus of any synod about the future of AI. Of course, a worldwide meeting would be ideal. We need to treat AI like atomic energy. It can be used for good, but it could also destroy us all. Ideally, an international framework for the development and control of AI should have already been written.
If you think that 19 men on airplanes who would kill almost 3000 Americans, not to mention the over 9000 deaths attributed to the toxic air at Ground Zero, should have been stopped before it occurred, then let’s take action now about AI. We KNOW that there are those in the world that want to use our own AI models to kill us and they won’t even have to get on a plane to do it.
The primary responsibility of any government is the safety of the governed. It is apparent that no one atop the big AI firms nor anyone in Washington, D.C. is remotely interested in making sure our own inventions cannot be used against us. Why?
Money. Everyone in AI is making too much of it and President Trump has no interest in curtailing their revenue streams, let alone regulating who gets to use the models. He just wants to “win” the AI race no matter the price our country and the world may have to pay.
The AI firms assume that they are sitting on the cusp of unlimited wealth generation with their technology. The problem is that they know, and now, so do we, that this technology can also be exploited by our enemies who seek to kill us. AI could be used to bring down the entire air traffic control system. It could shut down power generating grids. It could knock out every traffic light in Manhattan, scramble cell towers, and hurl orbiting satellites into downtown Los Angeles.
In light of the hand wringing about 9/11 during the 25th anniversary observance, instead of having ex-Presidents at Ground Zero or the current one speaking at the Pentagon, maybe all these leaders could get together with members of Congress and sit the AI CEOs down and threaten to nationalize their companies if they don’t instigate programs to guarantee the safety of the country. This is a matter of national security. Do it now, before it’s too late.
I hope I am not alone in thinking that the world may be on the cusp of another world war, because it seems pretty obvious to me. Iran has control of one oil lifeline, the Strait of Hormuz. The Houthis in Yemen have just secured another critical port in Mokha, Yemen at the bottleneck at the south end of the Red Sea. If this persists, the Iranians and their allies could easily resort to extortion to gain control of the oil market. Combine that with terrorists attempting to use AI to paralyze the United States or unleash a new pandemic and the threat to freedom and democracy is every bit as dangerous today as it once was in Nazi Germany and Imperial Japan in the 1930s. In fact, the current threats are even bigger. Nefarious use of AI could, indeed, lead to the extinction of mankind.
So, perhaps a case can be made for Trump’s War against Iran (to prevent a world war), BUT, only if he chooses to win it at all costs. Another president, I am thinking of FDR and Bush 41, would have created an alliance of free countries in America, Europe, Asia, and the Middle East who view the existential threat of Iran and AI as real. Together, an alliance of all the world’s free countries and their armies might be able to defeat Iran totally and coerce the AI CEO bros to gain greater control over who gets to use their products.
Not only don’t I see Mr. Trump doing this, I remain unaware of any current politician with the clarity of vision and the adult personality to pull this off. That’s too bad, because the threats to the United States and all its citizens are greater today than they were in early September 2001 when our leaders had a lapse in their imaginations and in their strategy to eliminate Al Qaeda.
Let’s not make that mistake again.
