Will A.I. Be a Creator or a Destroyer of Worlds?

"The advent of A.I. — artificial intelligence — is spurring curiosity and fear. Will A.I. be a creator or a destroyer of worlds?

In “Can We Have Pro-Worker A.I.? Choosing a Path of Machines in Service of Minds,” three economists at M.I.T., Daron Acemoglu, David Autor and Simon Johnson, look at this epochal innovation:

The private sector in the United States is currently pursuing a path for generative A.I. that emphasizes automation and the displacement of labor, along with intrusive workplace surveillance. As a result, disruptions could lead to a potential downward cascade in wage levels, as well as inefficient productivity gains.

Before the advent of artificial intelligence, automation was largely limited to blue-collar and office jobs using digital technologies while more complex and better paying jobs were left untouched because they require flexibility, judgment and common sense.

Now, Acemoglu, Autor and Johnson write, A.I. presents a direct threat to those high skill jobs: “A major focus of A.I. research is to attain human parity in a vast range of cognitive tasks and, more generally, to achieve ‘artificial general intelligence’ that fully mimics and then surpasses capabilities of the human mind.”

The three economists make the case that

There is no guarantee that the transformative capabilities of generative A.I. will be used for the betterment of work or workers. The bias of the tax code, of the private sector generally, and of the technology sector specifically, leans toward automation over augmentation.

But there are also potentially powerful A.I.-based tools that can be used to create new tasks, boosting expertise and productivity across a range of skills. To redirect A.I. development onto the human-complementary path requires changes in the direction of technological innovation, as well as in corporate norms and behavior. This needs to be backed up by the right priorities at the federal level and a broader public understanding of the stakes and the available choices. We know this is a tall order.

Tall is an understatement.

In an email elaborating on the A.I. paper, Acemoglu contended that artificial intelligence has the potential to improve employment prospects rather than undermine them:

It is quite possible to leverage generative A.I. as an informational tool that enables various different types of workers to get better at their jobs and perform more complex tasks. If we are able to do this, this would help create good, meaningful jobs, with wage growth potential, and may even reduce inequality. Think of a generative A.I. tools that helps electricians get much better at diagnosing complex problems and troubleshoot them effectively.

This, however, “is not where we are heading,” Acemoglu continued:

The preoccupation of the tech industry is still automation and more automation, and the monetization of data via digital ads. To turn generative A.I. pro-worker, we need a major course correction, and this is not something that’s going to happen by itself.

Acemoglu pointed out that unlike the regional trade shock after China entered the World Trade Association in 2001 that decimated manufacturing employment, “The kinds of tasks impacted by A.I. are much more broadly distributed in the population and also across regions.” In other words, A.I. threatens employment at virtually all levels of the economy, including well-paid jobs requiring complex cognitive capabilities.

Four technology specialists — Tyna Eloundou and Pamela Mishkin, both on the staff of OpenAI, together with Sam Manning, a research fellow at the Centre for the Governance of A.I., and Daniel Rock at the University of Pennsylvania — have provided a detailed case study on the employment effects of artificial intelligence in their 2023 paper “GPTs Are GPTs: an Early Look at the Labor Market Impact Potential of Large Language Models.”

“Around 80 percent of the U.S. work force could have at least 10 percent of their work tasks affected by the introduction of large language models,” Eloundou and her co-authors write, and “approximately 19 percent of workers may see at least 50 percent of their tasks impacted.”

Large language models have multiple and diverse uses, according to Eloundou and her colleagues, and “can process and produce various forms of sequential data, including assembly language, protein sequences and chess games, extending beyond natural.” In addition, these models “excel in diverse applications like translation, classification, creative writing, and code generation — capabilities that previously demanded specialized, task-specific models developed by expert engineers using domain-specific data.”

Eloundou and her co-authors note that well-paying, high-skill jobs are most vulnerable to replacement by artificial intelligence, but there are large differences for various occupations:

Our findings indicate that the importance of science and critical thinking skills are strongly negatively associated with exposure” — meaning vulnerable to replacement by artificial intelligence — “suggesting that occupations requiring these skills are less likely to be impacted by current L.L.M.s. Conversely, programming and writing skills show a strong positive association with exposure, implying that occupations involving these skills are more susceptible to being influenced by L.L.M.s.

Among the occupations Eloundou and her co-authors rank as most vulnerable are writers and authors, survey researchers, public relations specialists, interpreters and translators, web designers, financial analysts, court reporters, caption writers and tax preparers.

Just as there are those who emphasize the downsides of A.I., there are optimists who focus on the positive side.

In their 2023 paper, “Machines of Mind: the Case for an A.I.-Powered Productivity Boom, three economists, Martin Neil Baily of the Brookings Institution, Erik Brynjolfsson of Stanford and Anton Korinek of the University of Virginia contend that

Large language models such as ChatGPT are emerging as powerful tools that not only make workers more productive but also increase the rate of innovation, laying the foundation for a significant acceleration in economic growth. As a general purpose technology, A.I. will impact a wide array of industries, prompting investments in new skills, transforming business processes, and altering the nature of work.

Baily, Brynjolfsson and Korinek are not wide-eyed idealists.

“If labor can be replaced by machines across a wide range of tasks in the future,” they warn, “we may experience an A.I.-powered growth takeoff at the same time that the value of labor declines. This would present a significant challenge for our society. Moreover, artificial general intelligence may also impose large risks on humanity if not aligned with human objectives.”

These warnings, however, are issued in passing, in contrast to the work of Acemoglu, Autor and Johnson. The core focus of Baily, Brynjolfsson and Korinek is on the tremendous positive promise of artificial intelligence:

The potential of the most recent generation of A.I. systems is illustrated vividly by the viral uptake of ChatGPT, a large language model (LLM) that captured public attention by its ability to generate coherent and contextually appropriate text. This is not an innovation that is languishing in the basement. Its capabilities have already captivated hundreds of millions of users.

Other LLMs that were recently rolled out publicly include Google’s Bard and Anthropic’s Claude. But generative AI is not limited to text: in recent years, we have also seen generative AI systems that can create images, such as Midjourney, Stable Diffusion or DALL-E, and more recently multimodal systems that combine text, images, video, audio and even robotic functions.

These technologies are foundation models, which are vast systems based on deep neural networks that have been trained on massive amounts of data and can then be adapted to perform a wide range of different tasks. Because information and knowledge work dominate the U.S. economy, these machines of the mind will dramatically boost overall productivity.

Productivity, Baily and his co-authors go on to say, is “the primary determinant of our long-term prosperity and welfare.” They foresee artificial intelligence generating a virtuous circle, with productivity gains at its center: “If generative A.I. makes cognitive workers on average 30 percent more productive over a decade or two and cognitive work makes up about 60 percent of all value added in the economy, this amounts to an 18-percent increase in aggregate productivity and output.”

In addition, productivity growth will accelerate “innovation and thus future productivity growth. Cognitive workers not only produce current output but also invent new things, engage in discoveries, and generate the technological progress that boosts future productivity.”

How does this virtuous circle actually operate? It’s driven by the compounding of small annual gains into large multiyear improvements.

Baily, Brynjolfsson and Korinek observe that “if productivity growth was 2 percent and the cognitive labor that underpins productivity growth is 20 percent more productive, this would raise the growth rate of productivity by 20 percent to 2.4 percent,” a “barely noticeable” change:

But productivity growth compounds. After a decade, the described tiny increase in productivity growth would leave the economy 5 percent larger, and the growth would compound further every year thereafter. What’s more, if the acceleration applied to the growth rate of the growth rate, then, of course, growth would accelerate even more over time.

From a different vantage point, Autor sees the potential of a benefit for the expanded application of artificial intelligence. In his 2024 paper, “Applying A.I. to Rebuild Middle Class Jobs,” Autor argues that

The unique opportunity that A.I. offers to the labor market is to extend the relevance, reach, and value of human expertise.

Because of A.I.’s capacity to weave information and rules with acquired experience to support decision-making, it can be applied to enable a larger set of workers possessing complementary knowledge to perform some of the higher-stakes decision-making tasks that are currently arrogated to elite experts, e.g., medical care to doctors, document production to lawyers, software coding to computer engineers, and undergraduate education to professors.

My thesis is not a forecast but an argument about what is possible: A.I., if used well, can assist with restoring the middle-skill, middle-class heart of the U.S. labor market that has been hollowed out by automation and globalization.

There are fewer empirical data points in the study of the effects of artificial intelligence on the broad field of political competition, in comparison with the abundance of statistics and other kinds of information on jobs, economic growth and innovation. As a result, the scholarly analysis of A.I. and politics is a work in progress.

In his 2023 article “Artificial Intelligence and Democracy: A Conceptual Framework,” Andreas Jungherr, a political scientist at the University of Bamberg in Germany, maintains that “A.I. has begun to touch the very idea and practice of democracy.”

In the competition between democratic and autocratic states, Jungherr argues that artificial intelligence can help authoritarian leaders: “A.I. in autocracies creates an environment of permissive privacy regulation that provides developers and modelers with vast troves of data, allowing them to refine A.I.-enabled models of human behavior.”

Traditionally, Jungherr writes,

Democracies have been seen to be superior to autocracies due to their superior performance as information aggregators and processors. Free expression, a free press, and electorally channeled competition between factions provide democracies with structural mechanisms that surface information about society, the actions of bureaucracies, and the impact of policies. In contrast, autocracies restrict information flows by controlling speech, the media and political competition, leaving governments in the dark regarding local situations.

Artificial intelligence, Jungherr suggests, may enable “autocracies to overcome this disadvantage. The clearest example at present is China, which uses large-scale data collection and A.I. to support social planning and control — such as through its Social Credit System.”

Along these lines, artificial intelligence could provide authoritarian leaders access to the needs and views of their constituents, helping “autocracies increase their state capacities through A.I.-assisted governance and planning, increasing the quality of state-provided public services.”

If performed effectively and accurately, improved public services “might provide people living in autocracies with greater cultural, economic and health-related opportunities,” Jungherr writes, which, in turn, would encourage people to “see these benefits as a worthy trade-off with some individual freedoms, leading to strengthened public support for autocracies and state control.”

In examining the effect of artificial intelligence on politics, especially politics in this country, Bruce Schneier, a fellow at Harvard’s Berkman Klein Center for internet & Society and a lecturer at the Kennedy School, takes speculation to a new level.

In an essay that was published last week, “How A.I. Will Change Democracy,” Schneier writes:

A.I. can engage with voters, conduct polls and fundraise at a scale that humans cannot — for all sizes of elections. More interestingly, future politicians will largely be A.I.-driven. I don’t mean that AI will replace humans as politicians. But as A.I. starts to look and feel more human, our human politicians will start to look and feel more like A.I.

Artificial intelligence, Schneier believes, will shift power from executives — presidents and governors — to Congress and to state legislators:

Right now, laws tend to be general, with details to be worked out by a government agency. A.I. can allow legislators to propose, and then vote on, all of those details. That will change the balance of power between the legislative and the executive branches of government.

And finally, Schneier writes, taking his case a step further, “A.I. can eliminate the need for politicians.”

The system of representative democracy, he continues, “empowers elected officials to stand in for our collective preferences.” When the issues involved complex trade-offs, “we can only choose one of two — or maybe a few more — candidates to do that for us.”

Artificial intelligence, Schneier asserts, “can change this. We can imagine a personal A.I. directly participating in policy debates on our behalf, along with millions of other personal A.I.s, and coming to a consensus on policy.”

This consensus will be reached, Schneier maintains, by combining the data contained in devices he calls “personal A.I. assistants.”

These “assistants” according to Schneier, serve

as your advocate with others, and as a butler with you. This requires an intimacy greater than your search engine, email provider, cloud storage system, or phone. You’re going to want it with you 24/7, constantly training on everything you do. You will want it to know everything about you, so it can most effectively work on your behalf.

A.I. has revealed unfathomable vistas, as well as ungraspable, unrecognizable vulnerabilities — and the process has only just begun." [1]

1. Will A.I. Be a Creator or a Destroyer of Worlds?: Guest Essay. Edsall, Thomas B.  New York Times (Online) New York Times Company. Jun 5, 2024.