Don't Panic! Fear, Uncertainty, and AI

Introduction

In Douglas Adams' The Hitchhiker's Guide to the Galaxy, the words "DON'T PANIC" are printed in large, friendly letters on the cover of a guide for navigating the enormous, confusing, and occasionally dangerous universe. It's remarkably good advice. Panicking never makes a bad situation better, and it normally clouds our judgment.

Fear, Uncertainty, and Doubt

Companies use fear, uncertainty, and doubt (FUD) to discourage customers from switching to competing products by exaggerating the risks. But fear doesn't need to be manufactured. Real threats create genuine uncertainty, and uncertainty creates an opportunity for anyone who can offer a confident explanation of the problem and a plan to solve it.

Existential Threats

Threats presented as existential are particularly fertile ground for FUD. We can evaluate the risks of buying a house, driving a car, or changing jobs because we have experience with these decisions and ways to measure their consequences. But the human mind struggles to comprehend a threat to civilization itself. When the stakes are presented as the end of the world as we know it, normal questions about evidence, tradeoffs, and unintended consequences can start to look like distractions.

Climate Change

Al Gore's 2006 documentary An Inconvenient Truth won an Oscar and helped make climate change a mainstream political issue. Its persuasive power came partly from presenting a complicated scientific problem as an urgent, comprehensible story about a frightening future. But when a threat becomes a moral emergency, questioning a particular prediction or proposed solution can be mistaken for denying the problem itself.

COVID-19

COVID-19 brought this dynamic into everyday life. Public debates often collapsed several distinct questions into one: Was the virus dangerous? Would a particular restriction reduce that danger? What would the restriction cost? Could a less restrictive policy have achieved similar benefits at a lower cost? Proponents often equated questioning their policies' costs or effectiveness with downplaying the danger of the virus itself.

September 11

The September 11 attacks shattered Americans' sense of security. The resulting fear created an immediate demand for protection, and the public became willing to accept expansions of government surveillance and security powers that would have been politically unacceptable just days earlier. More than two decades later, the fear has faded, but much of the security apparatus created in response remains.

Nuclear Technology

Nuclear technology gave humanity the ability to destroy civilization, but it also offered a way to produce enormous amounts of reliable electricity with very low carbon emissions. Nuclear weapons established the technology's frightening reputation, and nuclear power plant accidents reinforced it. That fear probably led us to overestimate nuclear power's risks while underestimating the costs of going without it.

The Politics of Fear and Uncertainty

These examples share a pattern. A threat creates uncertainty, uncertainty generates fear, and fear creates demand for someone to take control. That someone might be a government, a regulator, a company, or an institution claiming special expertise. Once authority has been granted, the people and organizations holding it often have little incentive to give it back.

No conspiracy is required. Politicians benefit from appearing decisive, news organizations benefit from attention, and institutions benefit from funding and influence. Companies can benefit from regulations that make it harder for new competitors to enter their markets. These incentives don't mean the people involved are lying about the threat. But fear and uncertainty make us less critical of the solutions they propose.

The Fear of Artificial Intelligence

AI has all the ingredients for this pattern to repeat. The technology is advancing rapidly, its capabilities are difficult to predict, and its potential impact on society is enormous. That uncertainty creates an opening for anyone offering confident answers, while fear makes us eager to accept them.

Consider the prediction that AI will leave millions of people permanently unemployed. That's a possible future, not an argument for any particular economic policy. AI will eliminate some jobs, but it will certainly create new opportunities we can't yet imagine. The uncertainty is whether displaced workers will find new employment. Why commit to economic interventions before we know the scale of the problem?

Don't Panic

The more uncertain we are about AI's future, the more possibilities we ought to consider. Existential fear encourages the opposite: imagining a frightening possibility and demanding sweeping changes to prevent it. That's panic.

Douglas Adams' advice wasn't a promise that everything would turn out fine. It was a reminder of how to behave when you don't know whether it will.

Take AI seriously enough not to panic about it.