Suddenly, everyone is talking about AI and its potential to end humankind. Artificial-intelligence safety researcher Roman Yampolskiy has gained notoriety by saying there is a 99.9 percent chance that AI will destroy the human race, and former Anthropic researcher Jacob Coxon recently claimed that “people building AI believe it could kill all humans by the end of the decade.”
This brings back memories of the 1968 movie 2001: A Space Odyssey, especially the part where HAL 9000, the spacecraft’s supercomputer, malfunctions and acts on its own. When one of the astronauts carrying out the Jupiter mission in the film is locked out of the spaceship, he instructs HAL to open the pod bay doors, to which the supercomputer replies, “I’m sorry, Dave. I’m afraid I can’t do that.” The ultimate nightmare became reality: HAL had gone rogue and taken control of the spaceship!
Whether or not artificial intelligence could become an existential threat on its own is debatable, but the most dangerous aspect of AI is not AI itself. The greatest danger AI poses is the existence of evil people who would not hesitate to use a powerful weapon to control, injure, or destroy others.
Human nature did not suddenly become dangerous when the first computer was switched on. Human malevolence has existed since the beginning of time. Throughout history, every century has produced tyrants and who have had no qualms about terminating the lives of others. The reality is that even though technology changes, human nature does not.
Fire can warm a home or burn down a village. Chemistry can produce life-saving medicine or poison gas. Atomic science can generate electricity or vaporize a city. The morality of a tool is determined by the person controlling it.
AI is different because of the magnitude of the power it can give an individual. An evil man with a knife can kill a few people. An evil man with a nuclear weapon can kill millions. But an evil man with advanced AI could design lethal pathogens, penetrate networks, manipulate populations, control autonomous weapons, disable infrastructure, or start a war without exposing himself to danger.
AI can multiply intelligence, speed, reach, and destructive capability while reducing the manpower needed to carry out an evil plan. In the past, a dictator required armies of soldiers, spies, and bureaucrats to wreak death and destruction on entire nations, but today advanced AI could give a single individual the power to accomplish what once required an empire. Adolf Hitler, Mao Zedong, and Genghis Khan killed tens of millions of people, but imagine what they would have been able to accomplish had they been in possession of superintelligent AI.
Now, industry leaders like Dario Amodei, CEO of Anthropic, and Sam Altman, CEO of OpenAI, are calling for government regulation to protect mankind from an AI-generated Armageddon. Which sounds good in theory, but history has taught us that government itself is inherently evil and needs to be kept on a short leash at all times.
One of the most dangerous things about government regulation is that throughout history such regulations tend to protect giant corporations from competition, and those same giant corporations tend become major financial contributors to the same politicians who promote those regulations. History makes it clear that government regulation should always be viewed with suspicion.
Regardless, regulations are not foolproof when it comes to evil. Rules and regulations are obeyed by people who respect rules and regulations. Tyrannical governments, terrorists, criminals of all stripes, and political and religious zealots are likely to steal systems, remove restrictions, exploit security flaws, and build their own models. Does anyone really believe that countries like China, Russia, and Iran would abide by Western regulations—even if they agreed to do so?
Nor are respectable institutions immune to AI temptations. Some of history’s greatest crimes have been carried out by educated, well-dressed officials speaking the language of necessity, equality, security, and “the common good.” Power does not cease to be dangerous merely because it is exercised by someone in a pin-stripe suit and silk tie. (Think COVID pandemic.)
Unfortunately, there is no way to guarantee AI safety. Laws do not abolish murder. Treaties do not abolish war. Civilization is a never-ending effort to restrain the destructive minority without crushing the freedom of everyone else. Machines are new, but evil is ancient. And, unfortunately, nothing in human history gives us reason to believe that evil will ever cease to exist.
AI may cure diseases, increase prosperity, and eliminate much human drudgery, but the same power placed in the hands of people who dream of domination or mass destruction has the potential to kill millions of people. It’s a marriage made in Hell that will require our eternal vigilance.
Copyright © 2020 Robert Ringer
ROBERT RINGER is a New York Times #1 bestselling author who has appeared on numerous national radio and television shows, including The Tonight Show, Today, The Dennis Miller Show, Good Morning America, ABC Nightline, The Charlie Rose Show, as well as Fox News and Fox Business. To sign up for a free subscription to his mind-expanding daily insights, visit www.robertringer.com.
Help American Liberty PAC in our mission to elect conservatives and save our nation. Support – American Liberty PAC