Why We as a Society Are Thinking Badly about AI

If one lends an ear to the voices of today that are throwing their thoughts and feelings out into the society, what’s being said about AI is broken up into pieces that goes in a whole bunch of directions. It is utter chaos, utterly lacking in coherence as one group is cheering on AI’s contribution to great wealth, while another is seeing AI as an enemy that will rob them of their economic security, while another is envisioning how AI could become a weapon in the hands of gangsters and tyrants, while another is envisioning the great boon AI could be to human life.

That points to our problem with working effectively together as a society on this significant new development: an extraordinary intelligence that will clearly be able to do extraordinary things — for good and for ill.

The AI is so large a phenomenon that it can elevate a new corporation to being worth more than a trillion dollars, it can send some kinds of workers into unemployment, it can provide people with important assistance on a very wide spectrum of the things that people want to do, it can represent an environmental problem for today’s society, and it can conduct some important kinds of science very well. Each has a piece of a complex reality that’s important for us to understand. For us to have a constructive conversation, we need to be able to keep all those parts of the truth together in our minds.

Usually, we can deal with things in just two moves: judge it good or bad, then support or oppose it. Simple.”

But AI is not that kind of challenge. It simultaneously poses so many important dangers and offers so many important opportunities that a simple response will hardly suffice.

Challenges of this complexity don’t come along very often.

With COVID, we want to avoid it and defeat it. With climate change, we need to have the goal of eliminating the consumption of fossil fuels. With power that won’t let himself be restrained, we need to create “No Kings” rallies. We arrive at our proper destination in a couple of moves: this is bad, we want to stop it.

But with AI, we’ve got something that’s burst onto our pre-AI world, with all kinds of major implications that include both the destructive and the constructive.

AI is not yet our friend or our enemy– or rather it is already both.

Because different people are focused on different parts of that large and consequential reality, we are at odds like the proverbial blind men and the elephant: the ambitious tech bro who is getting ever richer is in touch with one part of the elephant; while the aspiring college student—resentful that the career they have been preparing for is now going to be done better and cheaper by AI—is in touch with another; while the neighborhoods facing the arrival of massive data centers encounter yet another part; and still another captures the attention of the defender of liberty worried about how AI could be used by tyrants.

But with different segments focused on different parts of the reality, each group ends up unable to understand the legitimate truths of the others. Instead of conversation, we get the mutual unintelligibility of the Tower of Babel.

The coherent conversation we need will require us to conceive of the whole elephant.

Elephant → Babel → Whole Elephant → Nuclear Example → Leadership → Self-interest → Whitewater

And if we’re going to be able to shape AI to be more our friend than our enemy, we’re going to have to have a conversation based on some amount of shared understanding that includes all the important truths—good and bad.

Gaining the understanding we need of this particular elephant—AI and the diversity of its potential consequences—is going to be a major historical challenge. Not only do we have to get beyond “for or against,” but there is enough that is unprecedented about this new kind of technology that we will also have to get beyond our habitual ways of thinking.

Within living memory, there was another technological breakthrough that rendered old ways of thinking obsolete. The dawn of the Age of Nuclear Weapons—which eventually led to an understanding of the need to avoid Mutual Assured Destruction—required those in charge of massive nuclear arsenals to move beyond the ways civilized societies had thought about war for millennia and find a path to mutual survival. (bring in concepts like deterrence structure?)

The unprecedented nature of AI, and its many-sided consequences, challenge us to transcend habitual ways of thinking. We are challenged not only to hold more ideas in our minds at once than we are accustomed to, but to think in new ways as well. And right now, humanity has only begun to explore what sorts of new thinking may be required of us in order to make beneficial use of this unprecedented kind of tool.

Unfortunately, we are at this moment hardly in a condition to rise to the extraordinary challenge posed by AI. At a moment when AI is demanding more of us than technological change usually does, we lack the leadership that might help frame for us the challenge we face. Indeed, the condition of our national leadership is such that there isn’t a single issue on which we, as a nation, can have a constructive conversation.

But we must prepare ourselves for the day when leadership might become part of the solution, rather than a perpetual source of our problems.

But leadership is not the only obstacle. There is another factor that must be overcome if our national conversation about AI is to become what we need it to be: the power of self-interest.

What comes easily to us is to see things in terms of our own costs and benefits, and seek to minimize the costs and maximize the benefits. If society is to think well and act wisely in dealing with this situation, each of us must also give weight to the question of what’s good for the human world, not only what’s best for me.

The conversation we need would require people to concern themselves with making things as they should be, not just ask, “What’s in it for me?” When history poses exceptional challenges, people often need more than their usual willingness to temper their pursuit of self-interest to serve the greater good. The collective challenge posed by World War II, for example, moved many people to make sacrifices on behalf of the nation.

Even with the best of leadership, and even with a revival of devotion to the greater good, this will be one of those times when history demands more of us than usual.

Most of the time, we simply row our canoe down the river of history, making whatever adjustments are required by the changing scenery and whatever branches may overhang the banks. But every now and then, history carries us into whitewater.

When the river changes, we need to change, too. When we come to whitewater, the relaxed attention and effort of the one paddling his canoe with the current will no longer suffice. We must give it everything we’ve got, lest we get dashed against the rocks.

AI has brought us to one of those stretches of river.

But there is another side to whitewater. There can also be exhilaration in meeting a great challenge well. There is a special satisfaction in finding our way through dangerous rapids—not merely surviving them, but steering rightly through them together.

If we navigate wisely, we can avoid dashing ourselves against the rocks of tyranny and manipulation and instead turn what might have been our enemy into a great friend—one that enhances people’s lives and helps societies find their way toward the futures they choose.