Making the A.I.-Climate connection

Anyone asked to identify the two biggest forces for change in the world today could do worse than choose artificial intelligence and climate change. Both are products of technology whose effects are only beginning to be felt, and the ultimate consequences of both will almost certainly be transformative in every sense of the word. Other than that, there hasn’t been much tying them together. Until now.

Welcome to Climate City, a label that a group of current and former data analysts and entrepreneurs has applied to Asheville, N.C. It might seem an unlikely spot for revolutionary thinking on such matters. We are, after all, nestled in Appalachia’s Blue Ridge Mountains in a state that seems hell-bent on taking the prize for most backward in the nation.

But Asheville is home to the National Centers for Environmental Information, the primary clearing house and analysis lab for the country’s — and increasingly much of the rest of the planet’s — climate data. It’s part of the National Oceanographic and Atmospheric Administration and until a couple of years ago had the less obfuscatory name of National Climatic Data Center. (Although the name was changed a couple of years ago, before Trump took office, I’m guessing NOAA brass realized that a less inviting target made sense given the science-denying, research-defunding druthers of the party that was and still is ostensiby running Congress.) A few hundred climate scientists work at the NCEI and a handful of other local university and government agencies that are still allowed to care about the climate.

James McMahon
James McMahon

So, to one familiar with the scientific culture of Asheville, it only makes sense that it would be the right coop to hatch the idea of marrying AI and climatology. And hatched it has been, by James McMahon, who for the last year has been the CEO of a unlikely non-profit called The Collider. The year-old organization offers physical space and social networking links to anyone who wants to be part of what is known as the embryonic “climate services” community. Friday morning meet-and-greets over coffee and pastries supplied by local bakeries are becoming a popular networking opportunity. This past week it played host to the annual meeting of the American Association of State Climatologists. (Who knew?)

Collider logoA proper definition of “climate services” is still a work in progress, but basically it refers to the provision of scientific advice and number-crunching for the benefit of private and public entities that need to worry about the effects of a changing climate. Think the city of Miami Beach, which is facing serious threats from sea-level rise. Think of any large corporation with physical assets. Think most of all about insurance companies. They all need people to tell them what to expect and when to expect it. What we have here is an emerging industry that, instead of just figuring how to forestall climate change, is using its expertise to profit from it.

Not that there’s anything wrong with that. Even climate scientists are entitled for make a buck.

McMahon, a former atmospheric chemist who worked on ozone depletion back in the day, has spent the last year watching, and helping, some of the smartest people navigate the high-tech consultancy start-up maze. More than a couple of senior scientists at NCEI have taken early retirement to pursue this path, and he’s decided to follow suit, after only a year as Collider chief.

His new company will draw on Silicon Valley brainpower and target Wall Street money in an effort to apply artificially intelligent systems to the problems posed by climate change. “I’ve been thinking about this for a long time, and just waiting for the market to be ready,” he told the Collider gang on his last day at the office. “It’s not, but I am.”

Kudos. I’m no expert in the subject, and have no idea if the plan merits investment. But anticipating what to expect and when to expect it is exactly what AI is all about. And humans haven’t exactly been doing all that great when it comes to solving the climate crisis. Maybe AI will be our salvation? “It will if I have anything to do about it,” McMahon told me while enjoying a slice of the cake served in his honor.

One could argue that humans have already come up with perfectly good solutions. The price of solar and wind power continues to plummet. Economic growth has mostly been decoupled from carbon emissions. And check out those Teslas. Isn’t the real obstacle political and corporate inertia? Yes. But maybe AI could help there, too. Indeed, maybe one of the hallmarks of genuine AI is how well it can be applied to socio-economic challenges.

Walking and chewing gum at the same time

Nature Climate Change has wandered into political science with a study from Stanford University. Seth Werfel’s examination of the “crowding-out” effect — the idea that humans have a tough time pursuing more than one strategy to solve a problem — is worth considering, even if its finding aren’t exactly earth-shattering.

The problem is laid out right off the top and requires no further explanation:

Household actions and government policies are both necessary to mitigate the effects of climate change. However, household behaviour may crowd out public support for government action by creating the perception of sufficient progress. … Further evidence suggests that the crowding-out effect may have been driven by an increase in the perceived importance of individual actions relative to government regulation and a decrease in the perceived issue importance of energy and environmental sustainability.

The results of the study’s survey, which involved 14,000 Japanese living in the wake of the 2011 Fukushima meltdown, show only a “marginal” effect, and some of the confidence intervals are pretty wide. But the conclusion is yes, people do tend to ease up on making political demands after they’ve taken a few steps to address the problem at home. No surprise there.

It may seem a bit trite. Activists are forever bemoaning the apparent inability of the citizenry to walk and chew gum at the same time. “You want me to buy new light bulbs (again!) AND call my congressman?” As Werfel summed up existing thinking among social theorists:

When people consider progress on a single sub-goal, additional actions toward achieving a superordinate goal are less likely to be pursued unless prior actions establish commitmentoward that goal.

But a couple of days after I came across the study, a piece in Slate magazine on a seemingly unrelated subject but with a similar theme got me thinking. In “Swim Lessons Won’t Keep Your Toddler From Drowning,” Melinda Wenner Moyer argues that too many parents treat swim lessons as sufficient when it comes to their responsibility for making sure their kids don’t drown. Instead of choosing supervision and lessons, they choose the latter alone and then return to their Instagram accounts. This despite the fact that drowning remains the leading cause of death from injuries among children between 1 and 4 years old.

The American Red Cross states that “the best thing you can do to help your family stay safe is to enroll in age-appropriate swim lessons,” which it starts offering at the tender age of 6 months. Yet the statistics are clear: Swim skills are simply not enough. Two-thirds of kids who drown, believe it or not, are excellent swimmers.

It’s not quite that simple, of course. Also playing a role is the tendency of many swim instructors to focus on technique and comfort levels instead of survival. The classes I took four decades ago were all big on staying alive (and I am happy to report that my offspring’s instructors this summer are still big on it), but apparently it’s not as top-of-mind as it used to be in all corners of the pool. It is another example, and one with even better stats to support it,  of a mindset that doesn’t allow for complexity. There’s no doubt that the crowding-out effect is real.

So what do we do about this latest addition to the long list of psychological obstacles (motivated reasoning, confirmation bias, plain old laziness, etc.) to getting people to care enough about anthropogenic climate trends? For me, it all goes back to the same place: People have to be taught to be critical thinkers in a democratic society. I know, I know. We don’t have time for pedagogical reform. Not when we have to get our carbon emissions down to zero two or three decades. But there’s no reason not to harp away at the idea that it is possible to act on parallel strategies simultaneously.

Meanwhile, if people are going to be intellectually lazy, then we can at least try to ease their stress by being more selective about what we ask them to worry about. We need to pay more attention to those things that pose real risks and stop fretting about every imagined catastrophe. When it comes to our kids,

When it comes to our kids, unsupervised play about the water’s edge is relatively hazardous. Unsupervised fort-building in the woods, not so much. We can apply the same thinking to environmental messaging. We all need to be responsible for our carbon footprints and we need to keep our members of Congress aware that we know how much of their campaign war chests is sourced from the fossil-fuel lobby. But maybe we don’t need to worry so much about nuclear reactors.