The biggest problem with the current generation of Artificial Intelligence is not that it will enslave us, or that the Magnificent 7 will lose trillions of dollars, or that many companies will be distracted by AI hype while competitors that stick to their knitting will obsolete their competitors’ technology.
Google has sunk tens of billions of dollars into developing its Gemini AI. Just days after its release, when it was asked to create pictures of World War II German soldiers, it showed a Black man and an Asian woman in Nazi uniforms. Google pulled Gemini from the market, because Gemini is dumb as a stump.
The real problem with the AI Hype Bubble is that companies that make safety-critical systems are caught up in this AI craze and are planning to put terribly defective AI systems in control of the things that our lives depend on: the power grid, bank payment systems, water treatment plants, self-driving cars, medical devices, hospital management systems, and traffic control.
We cannot survive without our critical infrastructure. If it were to break down or be held to ransom, food would not be transported to supermarkets; electricity and water would not be supplied to homes and offices. Within weeks families would huddle together in the cold and dark, without food or clean water.
It is imperative that defective dumb-as-a-stump AIs not be put in charge of the things that our lives depend on. The threat isn’t that Artificial Intelligence will rebel against us. The threat is that a terrorist, political rival, or enemy military force can defeat, confuse, hack, or degrade these stupid AIs and destroy or commandeer control of the systems. This is the AI doom that we must stop now.
Pushing forward the development of fundamental disruptive technology, like self-driving cars, is beyond the competence of most engineers. When they are confronted with a problem they can’t solve, they punt the job to AI, hoping that the AI can solve the problems they can’t solve. But it can’t because AI is dumber than a fifth grader.
We must not allow AI to be used in any system whose failure might result in loss of life, serious environmental damage, or economic collapse.
AI should be limited to what it does best: generating fiction, where truth has been abandoned for the purpose of entertainment.