From genetic engineering to quantum computing, incredible advances in technology are opening up policy landscapes and security challenges that were simply unforeseen even five years ago. His next decade would see smaller devices, larger networks, and anthropomorphic computers that augmented human thinking. Exchange Human thinking – literally beyond imagination and belief.
AI-powered doomsday predictions aimed at stirring up public anxiety have made headlines and become popular podcasts, but from the perspective of many PhD-level scientists and engineers, Borge's The predictions of life in the world are strangely exaggerated. One reason AI is gaining so much attention is that it, like satellite navigation and drug discovery, is nearly indistinguishable from magic. Large language models like Bard, Copilot, and ChatGPT sound like real people, which adds to their magical appeal. But they are full of mistakes and sweet-sounding illusions, and they are by no means free of mistakes.
Our obsession with AI diverts attention and energy from more immediate and transcendent threats to our society and global human progress. Dangerous disagreements are not based on moral values between humans and computers, but between humans and their ideological opponents. Irrefutable facts and valid (if false) opinions have been replaced by deliberately false ideas injected into our discourse like a powerful and addictive drug of delusion. Ta. If we can't agree on objective, reproducible scientific insights or a true historical record, how do we work together in the long-term best interests of our country and our planet? Facts that are now fed into AI computers There's nothing to filter the boring stuff out of.
The real danger is not the disaster caused by machines.it's bad human AI accelerates and enhances decision-making. There's a lot we think we know, from calculating financial risk to determining criminal recidivism. In the immortal words of Mark Twain, “It's not really that way.” Training computers based on discriminatory precedents is irresponsible at best and biased at worst. Repairing flawed ideologies in human memory or computer storage is extremely difficult, and it takes time to turn an ethical lens on both media.
In the real world, and in all of history before broadcast, new information and edicts spread very slowly, whether provable or not, sustainable or not. The worst ideas aimed at repression, exclusion, incitement, and subjugation were eventually eradicated, sometimes painfully, from the social system. Good ideas will take longer, including destroying bad ideas, but they will succeed in the end. As Mahatma Gandhi, Dr. Martin Luther King, Golda Meir, and Nelson Mandela reassure us, it is always a matter of time before harmful structures are defeated. The ultimate strategic intelligence is that true freedom is renewable, but not automatic. It requires discussion and generates criticism. AI can't do either of those things.
Another deficit and source of contamination is weaponized misinformation inserted into our nation's popular press and social media by foreign interests. These pathogens are fed into a training set that teaches the generation platform how and what to speak. AI has no ambition or judgment. This is truly advanced and impressive pattern recognition. Unless we are more careful and cautious, it will take years to eliminate toxic substances from our training sets and bring them in line with our expectations and laws.
Finally, global markets and (until recently) our national security depend on sophisticated components made in China. We taught them ourselves. Policymakers of both parties expected the Middle Kingdom to become a large market and friendly competitor. Rather, they are fierce commercial rivals and America's most worrisome military adversaries. They are already training almost 10 times the number of engineering students that we do, and soon he will be producing twice as many engineering Ph.D.s. The AI mismatch here is that it has more information than we do.
The clear danger now is not artificial intelligence. It's about honesty in training.
Just like real brains, AI only “learns” what we teach it. Today's computer models are susceptible to absorbing false ideas and disproved theories about science, history, economics, and philosophy. This is no different from schools that promote creationism, Holocaust denial, mercantilism, and theories of oppression disguised as real science. Ridiculous ideas are implanted in vast computer memories (now as big as a human brain), nullifying conclusions that sound real but cannot be independently verified, tracked, checked, or challenged. It creates discrimination. The real world effects are similar. These include spiritual superstitions, deep-seated suspicions, and manufactured conflicts.
AI has no imagination. It's a master mix of good and bad ideas that we've already considered. Sometimes the results are interesting, like a new chess move or a protein fold never seen before, and sometimes they're ridiculous. But focusing on AI itself won't solve anything. Instead, we need to focus on far better policies, as suggested by the brilliant title of the most influential computer science paper of the past decade. “All you need is attention.”
This discrepancy is not caused by machines, but only by human ingenuity. We need to be careful about listing “ingredients”, just as we already do for food, gasoline, medicine, and clothing. We need to make sure that we are teaching these machines things that are scientifically proven, socially consistent, and integrity tested for both accuracy and fairness.
Peter L. Levin is an adjunct senior fellow in the Technology and National Security Program at the Center for a New American Security and CEO of Amida Technology Solutions, Inc.
Copyright 2023 Nexstar Media Inc. All rights reserved. This material may not be published, broadcast, rewritten, or redistributed.
