AI Challenge Only Humans Can Solve | MIT News

The Dark Ages were not entirely dark. Advances in agriculture and building technology increased medieval wealth and led to a wave of cathedral construction in Europe. However, it was a time of deep inequality. Elites captured almost all economic interests. In Great Britain, when Canterbury Cathedral went up, peasants had no net increase in wealth between 1100 and 1300. Life expectancy varied around 25 years. Chronic malnutrition was widespread.

“We’ve struggled for a long time to share the prosperity,” says MIT professor Simon Johnson. “Every cathedral your parents dragged you to see in Europe is a symbol of despair and alienation made possible by higher productivity.”

At a glance, this might not be considered for life in 2023. But Johnson and his MIT colleague Daron Acemoglu, both economists, think it is. Technology drives economic progress. As innovations take place, one constant question is: Who benefits?

Scholars believe it refers to automation and artificial intelligence, which Acemoglu and Johnson’s Power and Progress: our 1,000-year struggle for technology and prosperity” is the focus of a new book published this week by PublicAffairs. In it, they examine who has benefited from past innovations and who may benefit from AI today, economically and politically.

“The book is about the choices we make with technology,” Johnson says. “It’s a very MIT kind of topic. But a lot of people feel that technology just falls on you and you have to live with it.”

AI could develop as a beneficial force, Johnson says. However, he adds: “A lot of algorithms are being created to try to replace humans as much as possible. We think this is completely wrong. The way we progress with technology is to make machines useful to people, not to displace them. We’ve had automation in the past, but with new human tasks and enough countervailing power in society.”

Today, AI is a tool of social control for some governments that also creates wealth for a small number of people, according to Acemoglu and Johnson. The current path of artificial intelligence is not good for the economy or democracy, and these two problems unfortunately reinforce each other,” they write.

Return to general prosperity.

Acemoglu and Johnson have collaborated before. In the early 2000s, with political scientist James Robinson, they produced influential articles on politics and economic progress. Acemoglu, an MIT institute professor, also co-authored with Robinson Why Nations Fail (2012), about political institutions and growth, and The Narrow Corridor (2019), which presents freedom as a never-secured outcome. social struggle.

Johnson, Ronald A. Kurtz Professor of Entrepreneurship at the MIT Sloan School of Management, wrote “13 Bankers” (2010) on financial reform and “Jump-Starting America” ​​(2019) with MIT economist Jonathan Gruber calling for more investment in scientific research.

In Power and Progress, the authors emphasize that technology has created remarkable long-term benefits. As they write, “we are much better off than our ancestors,” and “scientific and technological progress is an important part of that history.”

However, much suffering and oppression occurred over the long term, and not just in medieval times.

“During the Industrial Revolution, it was a 100-year struggle for workers to cut these huge productivity gains in textiles and railroads,” Johnson notes. Broader progress has been made through increased manpower and electoral management. When the U.S. economy grew spectacularly for three decades after World War II, the gains were widely shared, although not recently.

“We’re suggesting that we can get back on that path of shared prosperity, equipping everyone with technology and increasing productivity,” Johnson says. “We had all that in the post-war period. We can get it back, but not with our current form of machine intelligence. That, we believe, undermines prosperity in the United States and around the world.”

Call for ‘machine utility’, not ‘so much automation’

What do Acemoglu and Johnson think is flawed in artificial intelligence? First, they believe that AI development is too focused on simulating human intelligence. Scientists are skeptical of the idea that AI mirrors human thinking, even things like the chess program AlphaZero, which they see as more of a specialized instruction set.

Or, for example, image recognition programs. is it a husky or a wolf? – use large data sets of past human decisions to build predictive models. But they often depend on correlation (more likely than not, a hedgehog is in front of your house) and can’t repeat the same cues that people rely on. Researchers know this, of course, and continue to improve their tools. But Acemoglu and Robinson argue that many AI programs are less agile than the human mind and are suboptimal substitutes for it, even when AI is designed to replace human work.

Acemoglu, who has published numerous articles on automation and robots, calls these replacement tools “so many technologies.” A supermarket self-checkout machine does not add significant economic productivity; it simply transfers work to customers and wealth to shareholders. Or, for example, among more sophisticated AI tools, a customer service line that uses AI that doesn’t solve a given problem can frustrate people, causing them to vent when they reach a human and making the entire process less efficient.

Acemoglu and Johnson write that “neither traditional digital technologies nor artificial intelligence can perform critical tasks involving social interaction, adaptation, flexibility, and communication.”

Instead, growth economists favor technologies that create “marginal productivity” advantages that force firms to hire more workers. Rather than aiming to eliminate medical professionals like radiologists, a much-predicted AI development that hasn’t happened, Acemoglu and Johnson suggest AI tools can expand what home health care providers can do and make their services more valuable without cutting back on field workers. : .

“We believe that there is a fork in the road and it is not too late. AI is a very good opportunity to reassert the utility of the machine as a design philosophy,” says Johnson. “And find ways to put tools in the hands of workers, including low-wage workers.”

Definition of discussion

Another set of AI issues Acemoglu and Johnson worry about seeping directly into politics are video surveillance technologies, facial recognition tools, intensive data collection and AI-powered disinformation.

China is using artificial intelligence to create “social credit” scores for citizens, along with heavy surveillance while severely restricting freedom of speech. Elsewhere, social media platforms use algorithms to influence what users see; by emphasizing “engagement” above other priorities, they can spread harmful misinformation.

Indeed, Acemoglu and Johnson stress throughout Power and Progress that the use of artificial intelligence can create a self-perpetuating dynamic in which those who benefit economically can gain political influence and power at the expense of broader democratic participation. :

To change this trajectory, Acemoglu and Johnson advocate an extensive menu of policy responses, including ownership of Internet user data (a brainchild of technologist Yaron Lanier); tax reform that rewards employment more than automation; state support for a variety of high-tech research directions; repeal Section 230 of the Communications Decency Act of 1996, which protects online platforms from regulation or legal action based on their content; and a digital advertising tax (aimed at limiting the profitability of algorithm-based disinformation).

Johnson believes that people of all ideologies have incentives to support such measures;

Other scholars have praised “Strength and Progress.” Michael Sandel, Ann T. Professor of Government at Harvard University. and Robert M. anyone thinking about the fate of democracy in the digital age.”

For their part, Acemoglu and Johnson want to broaden the public discussion of AI beyond industry leaders, to abandon notions of AI’s inevitability and rethink human agency, social priorities, and economic possibilities.

“Debates about new technologies should focus not only on the brilliance of new products and algorithms, but also on whether they work for the people or against the people,” they write.

“We need these discussions,” Johnson says. “There is nothing inherent in technology. It’s under our control. Even if you think we can’t say no to new technologies, you can guide them and get better results from it if you talk about it.”

Source link