The Balancing Act: The artificial intelligence scare calmed with education

Many of us (humans) are afraid that AI (artificial intelligence) is going to take over the planet and kill off the human race. What B.S. AI needs food and that food comes in the form of one specific type of matter, the basic smallest matter on Earth, electrons, and a lot of them.

AI is powered by electricity and lots of it. A new data center needs millions of kiloWatt hours of the stuff — about a million continuous watts flowing constantly, never stopping at high voltages, like 220 or 440 volts. To imagine how much power this is, your home computer runs on 110 volts at perhaps less than 70 continuous watts, less than an old 75 watt incandescent lightbulb. My home, which uses a lot of air conditioning in 100-degree temperatures, uses about 2,000 kWh in 30 days. A small to medium AI data center might use at a minimum 100 mega watts of continuous power or 720,000 megawatts in a month. That is 360,000 times the power consumed by an average 2,500-square-foot home. In other words. a small- to medium-size AI data center uses the power of a community of a million people.

Shocking, isn’t it? Not really.

We have the current ability with “small,” “super safe” nuclear power to build and run 100,000 megawatt power plants and they should be made as part of the AI data center itself. Without getting into “super safe,” Small scale nuclear power cannot have a meltdown or thermal runaway by design.

But what about the “takeover” of human beings and their eventual demise. Stories of this can be summarized by perhaps two Hollywood movies: one from 1957 titled “Desk Set,” starring two great thespians, Spencer Tracy and Katharine Hepburn. The other film was “2001: A Space Odyssey” from 1968, which starred Keir Dullea, Gary Lockwood and Douglas Rain as the voice of the Hal 9000. It was a Stanley Kubrick film which also made famous the classical music piece, “Also sprach zarathustra.” It was the first film to really discuss AI although it was never referred to as such.

"Desk Set" was about a corporation installing an early computer (it had punch cards and such) for the purpose of allowing its research reference department and library to have quicker access to information. The research employees thought they were all going to lose their jobs and none did. But department productivity rose substantially and the two main characters fell in love. The film demonstrated the benefit of computers.

"2001: A Space Odyssey" was a bit different. It told the story of three astronauts on a space voyage to a moon of Jupiter to examine a monolith. The spacecraft had a “super computer” that spoke as a human called the Hal 9000. During the mission Hal got a mind of its own and killed off two of the astronauts during a period of hibernation and some how astronaut Dave survived. Dave had a decision, does he unplug Hal to potentially save his own life? He did.

"Space Odyssey" demonstrated out of the imagination of Kubrick and  screenwriter Arthur C. Clarke that a takeover could happen by a computer, and how it would be stopped. Simply remove the energy source for the computer via “unplugging” it. Remember during the film Hal begged Dave not to “unplug” him.

Is that the answer to the current AI question? The energy for all data centers is created and sustained by humans. If that remains the procedure and concept, then it would appear that humans will always have ultimate control. But what if an evil human decided he/she wanted to take over the world and let the super computer somehow kill off humans? A super computer could tell its government it is about to be attacked by nuclear missiles. Then a counter attack is launched and millions of people would die because of false information fed into the computer.

What if? Who has the power to unplug any “offending" AI data center?

I don’t exactly understand how a computer could directly kill us. Maybe in a hospital it is possible to turn off a heart/lung machine. But what about normal life? The microwave, a vacuum cleaner, the coffee maker or your own personal computer? Or could it be killer robots controlled by a super computer? Perhaps some reasonable regulation is required for robots?

It all remains for watching and study, but currently everyone seems to use the advantages of AI for obtaining the quickest and best possible answers to questions we have about law, medicine and about every other segment of life. We do need to be certain AI continues to give us the correct information, not sullied by political opinion. In other words, perhaps the best defense against AI is a good basic education to understand when AI becomes B.S.

Larry Weitzman is a former El Dorado County resident.