If AI thinks for us, democracy will suffer

While the Internet buried us in information, AI outsources analysis altogether

Sometime in the 18th century, the Enlightenment settled a contentious, centuries-old question: Where does knowledge come from? For millennia, many people had pointed to God. The early modern period brought a new consensus, namely that knowledge comes from human scientists working at the edge of what is known, armed with observation, mathematics, experimentation and logic.

As the historian Steven Shapin has said, this was a social as much as an intellectual revolution. Science started rewiring our shared value system around the idea that research and rigor are virtuous. We held ourselves and our highest institutions to the standard of scientific inquiry.

No political system reflects this shift better than democracy. The ideal of popular sovereignty rests on the principle that each of us is capable of comparing candidates or policy proposals and reaching our own informed conclusion.

Illustration: Mountain People

Today, we are on the cusp of a similar revolution. In September, one of OpenAI’s models took less than a week to solve a math problem that had stumped scientists for decades; and artificial intelligence (AI) systems are already identifying novel antibiotic candidates by searching chemical spaces too vast for any research team to explore.

In a growing number of fields, machines are autonomously producing new information. Humans are still posing the questions and evaluating the outputs, but the generative act — the creative leap across the unknown — is increasingly sparked by systems that do not think the way we do or really think at all.

As we saw with the Scientific Revolution, new information systems often lead to new epistemological and ethical systems. AI implicitly teaches us to value ease, speed and outsourcing, and these values do have their place in certain contexts. No human can parse as much data as quickly as an AI can, and in sectors such as medicine, speed often saves lives.

However, as anyone who has watched the Internet evolve in the past few years knows, moving fast can mean breaking things. We cannot allow our democratic institutions to become one of those things.

AI’s greatest skill is pattern recognition. We feed the algorithm a question or equation, and it compares inputs to identify an answer. What it cannot do is decide which questions are worth asking, which outcomes are worth pursuing, or what kind of future we should try to build. These are precisely the kinds of questions democracy demands we consider.

Many of the most important questions humans face cannot be outsourced. For example, which policies can mitigate political polarization? How should a liberal democracy balance surveillance and civil liberty? What does justice require of a society plagued by deepening inequality?

These are not problems that pattern recognition can solve, because none has a single solution. They require human judgement — however imperfect, contested, ever-evolving, and unsettled it might be — as well as context that only people in communities can recognize and feel. Whatever value system we adopt for our newest information age, we would need to account for AI’s inability to answer humanity’s most important questions.

We have been here before, and more recently than many people realize. Ever since the Internet became ubiquitous in our personal and professional lives, we have had to figure out how to metabolize more and more information, while protecting the truth. We have had to find ways to ensure that the digital economy serves the public interest, not just corporate interests and the bottom line.

Fortunately, litigants and courts have finally begun to hold the big social-media platforms liable for harms to children.

At the same time, digital technologies brought many benefits. The Digital Public Library of America has ensured access to books that are banned or unavailable in some parts of the country. The Internet Archive’s Wayback Machine allows us to track down an article on the Web from decades ago. Wikipedia has made a wealth of collective knowledge accessible for free. The Khan Academy has used digital technologies to democratize education.

These projects are testaments to the values we adopted centuries ago during the Scientific Revolution. Even as mis and disinformation became omnipresent online, we made a choice to protect truth and evidence from being buried, to make facts freely available and to welcome anyone who wanted to learn more.

AI requires us to make a similar choice. Whereas the Internet buried us in information — true and false, benign and malicious — AI offers to outsource analysis altogether.

We should resist the temptation. Answers, however derived, are not as valuable as insights (as mathematicians responding to OpenAI’s breakthroughs have pointed out). We should not value the product over the process of learning and discovery. Democracy demands that all citizens show up, whether at the ballot box, the school board meeting or the dinner table.

Our institutions, by design, cannot run on inertia. Democracy, like science, is powered by dissatisfied people always asking the next question about how we can be better, smarter, faster and freer. When we stop showing up or start substituting informed opinions for AI-generated answers, that engine stops.

Philanthropy can help. Along with nine other funders, the MacArthur Foundation is supporting Humanity AI, a US$500 million initiative to ensure that humanity remains at the center of the development of AI in the years to come. Two of the key areas of focus at the outset are democracy and education. These institutions must evolve to meet the moment, and AI itself needs to be designed and deployed always in ways that leave humans in the driver’s seat.

AI might ultimately empower us to create a better, smarter, safer world, but we would do well to remember that it was built on the backs of peer review and public research universities.

The products of the Scientific Revolution cannot supplant the process and the principles that created them. That process is precious, and it must be protected.

John Palfrey is president of the John D. and Catherine T. MacArthur Foundation.