Intelligence, artificial and other: our ruling class

The Council on Foreign Relations is widely regarded as a leading institution of the American ruling class. His colleagues design politics and his members, drawn from Wall Street, academia and elite journalism, rub shoulders with government ministers and even the occasional president. But its star has fallen with the demise of the old WASP establishment and its replacement of its two-party deliberative style with the crass rudeness of the present.

He can still attract a few famous names, even if the quality of the speech has declined somewhat. Last Monday, former Google CEO Eric Schmidt and former Theranos board member Henry Kissinger sat down (via Zoom) to discuss artificial intelligence (AI) – the subject of a new book which they wrote with Daniel Huttenlocher, the first dean of the Schwarzman College of Computing at MIT. The conversation, like the book, was a strange amalgamation of Schmidt’s technological enthusiasm and Kissinger’s central European gloom who largely accepted the breathless claims of AI promoters as fact.

The project began several years ago when Kissinger overheard a talk about a computer that had been programmed to play the extremely complex Go game. (Was this the first time he had heard of it?) Kissinger apparently began to worry about what all of this meant for the future of humanity, and wrote his concerns in a 2018 article in Atlantic. AI, Kissinger said, meant the end of the Enlightenment (which, to tell the truth, hasn’t looked too healthy for some time). “Human cognition is losing its personal character. People turn into data and data becomes king.

Big so true, as they say on the Internet. That machines can be so good at playing chess or Go may say more about these games than the potential of AI. Despite their complexity, the scope of these games is extremely limited and, for example, nothing compared to the seemingly mundane complexity of driving a car.

I’ve been following the advancement of AI for twenty years or so and the story has always been the same: a handful of successful examples portend a vast payoff that’s still just around the corner, but no never quite happens. Claims for autonomous vehicles are particularly grandiose at the moment. Not a day goes by that Elon Musk does not extol the autonomous driving skills of his Tesla. The reality is quite different.

“Fully autonomous driving” is a long way off, as CNN reporter Michael Ballaban showed just weeks ago with his attempt to let a Tesla drive him safely along a dangerous passage in the Atlantic. Brooklyn Avenue, a densely populated and chaotic stretch of road. whose perils I fear every time I sail it. Only his intervention kept the car from driving him into an oncoming UPS truck, and that was just one of many near misses. The Ballaban misadventures come three years after an autonomous Uber killed a pedestrian in Arizona. It took several fatal seconds for the software in that car to figure out that what it initially thought was a bicycle was actually a person. He finally decided to brake far too late. Obviously, the software hasn’t progressed much since.

In their book, Kissinger, Schmidt, and Huttenlocher do a lot of GPT-3, the latest iteration of an AI project that can produce words that look a lot like human speech. He has many dazzling abilities and can even write believable prose like this Guardian article. Well, not exactly. The machine spat eight different attempts, which the editors turned into a publishable article by picking “the best parts of each.” And GPT-3’s answers to simple questions are often silly, untrue, and even racist, but these embarrassing parts are rarely featured in public protests, whether as a normal tech boost or out of a desire to impress capital investors- risk.

Just for a moment, let’s give the point that AI is something big that is changing the way we live. Schmidt and especially Kissinger are worried about what it means to be human. (It’s strange when the architect of the secret bombing of Cambodia becomes the humanist of the program, but such is the policy of elite organizations.) Over the next 15 years, Schmidt argues, computers will increasingly establish their own program, exploring ways and producing results beyond the intent or comprehension of their human programmers. What is it going to do to our sense of ourselves, Schmidt asked, “if we are no longer the smartest person when it comes to intelligence?”

An answer might be, “Well, maybe don’t let them go?” But the authors will have none of that. “Once AI’s performance surpasses that of humans for a given task, failing to apply AI, at least as a complement to human efforts, can increasingly appear perverse, if not negligent,” they say. Are we going to delegate our warfare capabilities to the machines, not only to guide the weapons to their targets, but also to decide whether to attack first? Schmidt apparently thinks so, although he recognizes that there are some complexities. “So you’re in a war and the computer correctly calculates that in order to win the war you have to let your aircraft carrier sink, which would kill 5,000 people, or whatever….” Would a human make this decision? Almost certainly not. Would the computer be ready to do it? Absoutely.”

In a review of the book, Marc Rotenberg complains that all of Kissinger’s skepticism on the part of Atlantic the article disappears in the techno-euphoria of Schmidt and Huttenlocher. That is not exactly correct; the book presents the occasional irruption of Kissingerian sadness and heavy reflections on Descartes and Spengler. But it is above all true, and he is right to distinguish this “perverse or even negligent” passage as symptomatic of the “unchallenged assertion of inevitability” of the two technological authors.

Oddly, however, at the CFR event, Rotenberg asked how AI could be designed to “strengthen democratic principles.” Democracy is meant to be about transparent deliberation and debate, for which AI doesn’t seem like a good solution. In response, Schmidt first referred to Kissinger, a man whose democratic sensibilities were nicely summed up in a remark he made in the run-up to the coup against democratically elected Chilean President Salvador Allende. : “I don’t see why we have to stand up and watch a country become communist because of the irresponsibility of its own people. On the question of the democratization of AI, Kissinger admitted having no idea: “I think it needs to be studied, but I don’t know how to approach it yet. Resuming, Schmidt acknowledged that we needed to think about the subject, but quickly turned to criticism of Europe’s penchant for regulation.

Reading the book or listening to the CFR session leaves a little wiser about the technical realities of AI or its political implications. Kissinger and Schmidt are right that we need to talk about these things, but an Imperial master architect and what software engineer and AI skeptic Dwayne Monroe calls “a hype man for a super busy advertising company.” aren’t the ones leading the conversation (although Schmidt has left Google, he still owns around 1% of the company’s shares).

In the words of computer scientist Jonathan Bennett, “the real danger is not that AI will acquire superhuman powers but rather that we are convinced of it, and we will cede to it important tasks of reasoning than it is. not able to accomplish. As the sophistication of AI salon tricks increases, this possibility increases with it.

AI projects like GPT-3, as Luciano Floridi and Massimo Chiriatti assert, should be seen as tools that humans can use, and not as something that seriously resembles human intelligence: “GPT-3 is an extraordinary technology, but also intelligent, conscious, intelligent, aware, perceptive, perceptive, sensitive and sensitive (etc.) like an old typewriter. Maintaining such a perspective, rather than giving in to the inevitable domination of our silicon masters, is the starting point.

One consolation: it was satisfying to hear Kissinger the day after a socialist was elected president of Chile, in part by promising to undo the five decades of reaction launched by the coup that Kissinger had planned a lot. . I wanted to ask him the question, but only members of the CFR are allowed to ask questions during these events. Journalists are there only as passive receptacles.


Source link

Comments are closed.