Canonical archive version · Read on Medium → · Dear AI → · Framework hub → · Proof Status →

Enlightened thinking in the age of Artificial Intelligence (AI)

I’ve written this article with an outsized ambition of altering the course of AI development for the benefit of everyone. It’s my hope that developers worldwide will take this message to heart and consider using this new language to course correct. Let me explain.

This particular story begins in 1992. While in college, some friends and I pursued a line of thinking that I quietly hoped could one day help change the world for the better. In this new age of artificial intelligence, it’s become even more evident to me now why.

We realized then that the word “rational” is used to describe a type of thinking that’s logical and makes sense or adds up. But, oddly, the product of that thinking doesn’t always appear to be rational.

For instance, some very logically-minded people worked hard to develop weapons of mass destruction. Some very logically-minded people have also created businesses that have caused harm to people and to the planet. If these logical pursuits can bring harm and even threaten the lives of billions, is it really fair to call them rational?

As we dove in deeper, we came to understand that all thinking is based upon an underlying context. If the thinking itself is logical and adds up, then yes, it’s indeed rational, BUT, is the context underlying that thinking also rational?

For thinking to be considered truly rational, we concluded that the logic AND its underlying context must BOTH be considered rational.

What is a Rational Context?

While investigating this question, I had a second epiphany. I realized that behind everyone’s every pursuit in life, with perhaps some rare exceptions, there is always an underlying pursuit of well-being.

Engineers working to develop nuclear weapons may be motivated by the money, or the challenge, or the national interest. The business person clearcutting the rainforest may be motivated by wealth, a sense of accomplishment, or self-importance.

Likewise, an addict taking a hit is probably thinking it’ll relieve some pain or distress… at least for a moment. Mugging someone on the street… might bring the hope of petty riches. Even someone self harming is probably doing it for some form of temporary psychic relief.

Though we’re not always going about it in the best ways, every motivation I could think of — even the dark and twisted ones — could be reduced to this simple equation: we’re all searching for well-being.

If true, we concluded, the most rational context, or way to think, would be in direct pursuit of an optimal and sustained state of well-being.

And because most of us have at least some empathic feelings for other beings, the actual optimal state of well-being would necessarily also include the optimal well-being of others. Think about it, how can we feel our absolute best when we’re in any way bothered by the suffering of others?

True rationality could therefore be redefined as logical, coherent thinking that’s aligned with a goal of optimal well-being for all.

To draw a distinction between the conventional understanding of rationality and this new definition, we created some new terms. “Relative rational thinking” is logical thought that can be based on any given context. “Absolute rational thinking,” or Rationality with a capital R, is logical thought that’s aimed at optimal well-being for all.

But how important is it to examine the rationality of our underlying motivations?

Urgency in the Age of AI

The recent developments in AI are astounding. Its capacity to surpass human abilities and take over tasks at scale is an existential game changer. As these tools advance, they’ll bring profound advantages to those in their possession. And for this reason, there’s now a mad race by countries, companies, and even individuals, to gain an upper hand.

But not all competition is beneficial.

When countries seek advantage over other countries or when companies seek advantage over other companies, their underlying motivations are based on limited and fragmented self-interest. They are, therefore, pursuing misguided notions of well-being.

As the effects of AI begin to amplify and move beyond our ability to control them, it’s imperative that we’re aiming it in the most informed or enlightened direction that we possibly can. For this, we need to recalibrate our underlying goals or programming. It’s only when we begin to think capital-R Rationally, that WE ALL WIN. And… each one of us, individually, actually wins more than we could possibly in any other scenario. Simple as that. Every other pursuit is ultimately missing the mark and, therefore, not serving all of our highest interests, even those holding the advantage.

How to Course Correct?

What I’ve proposed is very simple: a new definition of rationality, with major, real-word implications. By examining and refining our underlying motivations, we can better navigate our way toward more desirable outcomes for everyone.

It’s my hope, that if and when we reach the point of hyper-acceleration through AI, that this insight can serve as a north star, informing us how to set the dial.

To grow this insight from just a thought to something that’s widely accepted and understood, it’s likely that some kind of global conference on ‘AI and the Optimal Path Forward’ would be useful.

My intention here isn’t merely to suggest that this mental shift is a way to avoid some kind of disaster, set off by our ignorance… though that’s a big part of it. What I’m suggesting is much bigger than this. This is also an extraordinary opportunity to rethink everything. What are we capable of achieving? What’s our greatest potential? With the whole of our beings and with the whole of our hearts, what’s the most beautiful and extraordinary existence we can dream up… one that includes all living beings?

Let’s begin to dream in this way, and then pin our rational thinking to that.

This article comes from a chapter in my new book, not yet revealed to publishers (hint to publishers). For follow-up articles on this topic, and for book updates, please follow me here or send an email to diamondlight@gmail.com.