A computer with its own intelligence speaks to humans and gives advice in a remarkably gentle voice. “I can see you’re really upset about this,” says the machine. “Honestly, I think you should sit calmly. Take some stress pills and think about it.”
Of course, this scene is from the 1968 sci-fi thriller, one of the greatest films of the 20th century. 2001: A Space OdysseyAnd perhaps the most famous computer name ever is HAL, short for Heuristically Programmed Algorithmic Computer. The astronaut who’s big on the screen when HAL throbs and calls out is named Dave, but he’s really a stand-in for all of us.
Proof of this is that we are nearing a moment when the foresight and terrifying vision of the film’s creators, Stanley Kubrick and author Arthur C. Clarke, will finally come true, or even turn around. nothing more than the fact that To the objective reality in front of you.
the film 2001 The work can be interpreted in many ways, but at its core, it’s a technological dystopian tale in which an all-powerful, sentient computer that has never made an error goes on a horrifying rampage. Start, hijack a manned spaceship, and kill the crew cold-heartedly. . Two of the astronauts board the space pod unobtrusively while planning to shut down HAL. One of them soon strays into the void and is murdered by a computer. Dave, the lone survivor, finds himself trapped outside the spaceship and orders HAL to open the door and return him to the spaceship.
“I’m sorry Dave, I can’t do that,” HAL replied in a surprisingly smooth voice.
“What’s wrong?” Dave asks.
“I’m sure you know what the problem is as well as I do,” HAL replies. “This mission is too important to me, and I won’t allow you to risk it.”
Dave and the last surviving crewmate were very careful not to let HAL hear of their plans, but through a small window in the pod, a computer that could read their lips from a distance. I didn’t expect the ability. Dave eventually returns to the mothership and successfully disables HAL. But perhaps the scariest part of the film is the idea that if Kubrick and Clark were making masterpieces today, they wouldn’t have thought that humans could pull themselves out of this predicament so well. do not have.
Over the last few months, I’ve been following developments in the artificial intelligence (AI) world and the most popular representations such as ChatGPT. And the impression these developments have left on me is that no matter how quickly we imagine this technological landscape may change, our vision just keeps falling behind. I was placed in HAL’s heart, 2001But an item I stumbled across just a week ago made the film’s lip reading already seem hopelessly outdated. In the brave new world of AI, scientists and engineers have already perfected how to read people’s brain activity—our thoughts—by simply monitoring the electrical signals we all emit, so we can’t rely on such indirect You may soon no longer need to worry about how. Release.
In recent weeks, there have been a flurry of warnings from those who know the field best and from so-called curious bystanders, people like myself, suggesting that humanity has unleashed a monster. . You can control it. How AI could soon be used to do things like rigged elections, commit massive financial fraud, help unlock new weapon systems, and promote the worst kinds of political propaganda. Beyond a series of warnings about what is going on, we are now inundated with reports of AI abuses. The impact is now being felt through applications that have already been introduced into the workplace and consumer markets.
Computer giant IBM, for example, has cut thousands of jobs driven by its power, raising concerns that the imminent threat of AI could pose jobs in countless other industries, such as Hollywood writers. causing anxiety and confusion. Meanwhile, the off-the-shelf technology that you and I can already buy and use is essentially creating the ability to create convincing virtual clones of ourselves.
Perhaps the most famous of all the many warnings so far is from AI pioneer Jeffrey Hinton, who recently resigned from Google. In an interview with BBC News, terminatorHe warned of endless disaster scenarios that could be caused by style automata, potentially rogue doomsday machines, and AI technology at the hands of what is often called the villain.
What we’re seeing now is something like GPT4. It goes far beyond the amount of general knowledge one has. I’m not very good at reasoning, but I can already do simple reasoning. And given the speed of progress, the situation is expected to improve rather quickly, so we should worry about it. As far as I know, they are not smarter than us at the moment, but I think they may be soon.
The next speaker in our BBC News corner was Junaid Mveen, a research mathematician and expert on human-AI interfaces. Towards the end of the conversation, he concludes with a dark conclusion:[Hinton’s] My own alert seems to arrive quite late in the day. And then there’s the feeling that perhaps the genie has come out of the bottle, with no real explanation for what happens next. “
These days, editorials in major Western newspapers are filled with sounding but vague advice on how to deal with this potential threat. This week, for example financial times The editorial board called on the tech industry to “agree and implement some common principles of transparency, accountability and fairness,” and called the magazine “critical areas of life and death, such as health care and the justice system.” He added that what he calls a computer algorithm is being used. Prior government approval is required.
But I have to say this is very creepy. Let me tell you why. Google is the best place to explain why such a suggestion seems so impossible. The hugely profitable search engine business has long quietly worked to figure out how to integrate AI into its products, but worries the risks are not yet fully understood. and withheld. Then Microsoft announced a powerful new product of its own called Edge. It is embedding ChatGPT (in which Microsoft has invested) into its own browser and search products, whose power and novelty will allow Microsoft to make dramatic gains in a market dominated by Google. I was betting. has long been effectively single-handed.
The point derived from this could be said to be principle number one. That is, profit usually trumps prudence when it suddenly looks like a big profit. Google adheres to this principle and, as you can imagine, has followed Microsoft enthusiastically, overcoming much of its previous caution.
Principle 2 extends beyond corporate greed, which many have taught us to be the positive force in capitalism, to the equally powerful realm of logic of the nation-state. Is it possible to imagine a world where China doesn’t try to match the recent advances that big US companies have made in AI applications? Weeks are rushing to roll out a product that rivals Google and Microsoft. And this isn’t just about her two superpowers in the world. Others who want to catch up, or at least catch up, will inevitably throw their hats into the ring in a contest that fairly reflects nuclear proliferation.
I’m not going to be totally hopeless here, but to predict how this situation will turn out when another technology looming on the horizon — quantum computing — finally comes to maturity. is almost impossible. Combining this with AI will exponentially increase the challenges humanity faces in keeping machines safe.
But my deepest concern goes back a long time ago, when some visionaries already began to foresee all this happening.movie 2001: A Space Odyssey Two years later, it was followed by another cultural blockbuster, now largely forgotten. In 1970, in the book future shock The book was a huge bestseller at the time, but author Alvin Toffler, along with his uncredited collaborator and wife Heidi, argued that some of the specific technical details had eluded their imagination for over 50 years. also foresaw the whole human dilemma I described here. It seems obvious.
In his often clumsily worded book, Toffler spells out the warning quite plainly—for my money, I’m better than a magazine editorial writer. financial timesand at the same point:
[T]Ecological questions can no longer be answered in technical terms alone. Those are political questions. In fact, they affect us more deeply than most of the superficial political issues that occupy us today. This is why we cannot continue to make technical decisions the traditional way. We cannot allow them to be haphazardly created independently of each other. We cannot allow them to be determined solely by short-term economic considerations. We cannot allow these to take place in a policy vacuum. And responsibility for such decisions cannot be casually delegated to businessmen, scientists, engineers, or managers who are unaware of the grave consequences of their actions.
But that’s exactly what we seem to be doing.
Another thinker helps to understand why, but his conclusions are by no means encouraging.in his book Looking Like a Nation: How Certain Plans to Improve the Human Condition Failed Yale University political scientist James C. Scott writes about the tendency of humans to make “heroic assumptions” about the power of cleverness to find solutions to any problem they face. The book explores what Scott calls “advanced modernism,” the use of science and theory to order and regularize the social world, and the use of theories of the future to recreate the present. Scott defines advanced modernism in these terms:
It is best thought of as a powerful, one might even say muscle-bound version of the belief in scientific and technological progress associated with the industrialization of Western Europe and North America from about 1830 until World War I. It is It is the supremacy of continuous linear progress, the development of scientific and technical knowledge, the expansion of production, the rational design of social order, the ever-increasing satisfaction of human needs, and, above all, the increasing control over nature. confidence.
But the lessons to be learned from AI and the formidable upcoming breakthroughs such as quantum computing are that as we reach the point where only the most unnatural thing to humans can save us: humility. It means that there may be.
