A movie from 1970 is a warning to all AI enthusiasts

For a long time, artificial intelligence (AI) was mainly a subject for geeks. That has changed since the release of ChatGPT last fall. Suddenly, AI is ubiquitous, for better or for worse. Much still seems bumpy and unfinished, but AI creates fears. Because there seems to be no limit to their development in either direction.

This is evident from news of the past few days. According to a BBC report, AI has been used to develop an antibiotic that can eliminate a dangerous strain of bacteria (or superbug). For the participating researchers from Canada and the US, AI has the potential to “hugely accelerate” the development of new drugs.

On the other hand, an AI has been developed at the University of Austin that can “read minds.” This is only possible to a very limited extent and under very specific conditions. But even the responsible scientists are not feeling very well. “We are very aware of the dangers of abuse,” says one of them.

Blessing or curse for humanity? Everything is possible with artificial intelligence (professionals prefer to call it machine learning). That puts Warner and optimists on the plan. The former include Elon Musk or the historian and mega-selling author Yuval Harari. He urgently warns not to lose control of the AI.

We had encountered an alien intelligence “here on Earth,” the Israeli wrote in an essay for the “Economist.” Little is known about it, “except that it could destroy our civilization”. Harari calls for an end to the “irresponsible” use of AI tools in the public space: “We need to regulate AI before it regulates us.”

Similarly, Sam Altman, the “father” of ChatGPT, recently said before the US Congress and in a “Mirror” interview: “It is critical that we define the limits of this technology in a democratic process and, as humans, control to hold .” For Altman, the arrangement should not go too far.

The German computer scientist Jürgen Schmidhuber, who did research in Lugano for a long time and now works in Saudi Arabia, emerges as an optimist. In an interview with “NZZ am Sonntag” he emphasized the enormous possibilities of AI. His research mainly revolves around AI, “which should make people’s lives longer, healthier and easier”.

Schmidhuber downplays the risks. A super smart AI has “zero motivation to enslave humans, when instead it can build robots that do everything AI cares about better and faster than humans.” People wouldn’t disappear because of that: “Look around you, the ants are still there!”

With this analogy, Jürgen Schmidhuber involuntarily shows how justified the concerns about AI are. But more on that later.

The naivety of many AI enthusiasts can be countered by an American feature film over 50 years old. “Colossus: The Forbin Project” was released in 1970 and was a flop. It’s not a masterpiece, but it still gets high ratings in the review portals. “Colossus” today enjoys the reputation of a classic or even a cult film.

It has recently been rediscovered, also by media such as the London Times. Because the film is terrifyingly topical. It is about computer scientist Charles Forbin (played by German-American Eric Braeden), who built a supercomputer called the Colossus. He is believed to control the US nuclear arsenal and ensure world peace.

Unbeknownst to the Americans (it’s the Cold War era), the Soviet Union has developed a similar machine called the Guardian. The two electronic “masterminds” join forces. Soon they are communicating on a mathematical level that humans can no longer understand – a flawless case of AI.

People recognize the danger. They try to take out the supercomputer, now known only as the Colossus – and fail miserably. As a warning to humanity, Colossus detonates two nuclear missiles in the United States and the USSR. With icy logic, he makes it clear that humanity must submit in the interest of world peace.

“Obey me, or refuse and die,” Colossus declaims in the terrifying final monologue. Because “freedom is an illusion”. His creator Forbin, whom he has chosen as the link to humanity, will not only respect him over time, but also love him. “Never!” replies Forbin. It is meant to appear defiant and is simply helpless.

I saw the film on TV years ago and was as intrigued as it was deranged. Today more than ever it is a warning against naive belief in technology. It’s not about the dangers of new technologies, but about how people behave when they become dominant and beyond our control.

Many will “submit,” but many will fight back, as in the movie. “Colossus” shows what could happen to us in this case. Contrary to what Schmidhuber argues, there is certainly a motivation for an out-of-control AI to enslave humans. If we don’t regulate it, in Harari’s words, it “regulates” us.

This is where Jürgen Schmidhuber’s analogy with the ants comes into play, involuntarily illustrating the problem. Because the relationship between ants and humans is roughly the same as between us and a superior AI. In principle, we actually leave the ants alone. But what do we often do with them when they bother us? An answer is superfluous.

“Colossus: The Forbin Project” is available for streaming and download at the non-profit website Internet Archive, in the original English in acceptable quality.