Over the past few years, Silicon Valley has steadily released into the world a raft of computer programs with capabilities that seem ripped from the pages of science fiction. So-called artificial intelligence (AI) can now beat us at games of strategy, predict the three-dimensional structure of proteins, generate sparkling prose on almost any topic in the style of nearly any writer who ever lived, and produce gorgeous photographs and paintings in response to brief, simple instructions.
These developments offer incredible promise. AI may not only rid our lives of everyday drudgery, giving us the time to do the things that give us joy while bots drive our cars, respond to our emails, and write our code, but also help us to realize our full potential as human beings. AI-based instructors trained on our individual learning styles could take us to new intellectual heights; AI-based doctors and therapists could tend to our physical and mental health; and AI-based creative tools could unleash our inner artists and musicians.
But with the promise comes a fair measure of dread. If AI can replace workers—from computer programmers to drivers to teachers, physicians, and artists—how will the displaced earn their living and fill their lives with meaning? Even more worrying is the fact that the new capacities with which AI promises to endow us can just as easily be put to evil ends as to good ones. Autocrats can better monitor and control their citizens; criminals can better prey on the unsuspecting; and enemies of freedom can better sow chaos, mislead voters, and polarize societies.
Bad intentions, however, are not necessary for AI to have bad effects: Since most of these programs are trained using real-world data, they have the potential to reproduce real-world dysfunctions. If, for example, a recruitment AI is trained on millions of decisions made by millions of actual hiring managers, and hiring managers tend to discriminate against minorities, then so too will the AI, but now with a veneer of objectivity. And that is not the worst of it. With the rapid pace of AI development, we may soon find ourselves in the same unenviable position as some of our hominid ancestors when Homo sapiens appeared—sharing the planet with intelligences far greater than our own and continuing to exist only at their pleasure.
In the pages that follow, we bring together a group of leading thinkers, experts, and technologists to explore the challenges that artificial intelligence poses for humanity, and how democratic institutions can be marshaled to help meet those challenges. Will AI shower wealth on everyone, or will it merely deepen the inequalities that everywhere undermine democracy? Will it provide citizens in autocracies with new tools in the fight for democracy, or will it just give autocrats new ways to fend off democratic challengers? If AI becomes smarter than we are, will it put an end to mankind, merely put us out of work, or enable humanity to flourish as never before? And can democracy make the more benign outcomes more likely than the malign ones, or is the rule of the people ill-suited to a world in which earth-shattering technologies can be unleashed at breakneck speed with only the dimmest awareness and understanding on the part of citizens and their representatives? These are hard questions, but they are fast becoming the questions of our time.
Copyright © 2023 National Endowment for Democracy and Johns Hopkins University Press