One of the most fundamental ideas in my work as a management consultant and coach is that human beings, all of us, are continuously creating our ‘reality’ through conversations and coordinating action. Our present reality is a product of actions taken in the past. Our future will similarly be a function of the actions that you, I, and others are taking right now.

Sometimes the future we create is intentional. Sometimes it is not. The future, even if it maintains the status quo, is, however, always being created moment to moment by human beings speaking and listening in conversations.

But what is action?

And if we live our lives in conversation, where is the action?

If you accept the premise that consciousness—your awareness, perceptions, thinking, and even your feelings and emotions—is occurring in language, then you must ask yourself what about action? Does action also occur in language?

To paraphrase one of my favorite quotations by Martin Heidegger:

We have not examined the nature of action decisively enough: we believe action is about causing results and then we repeat or modify the action based on what we think of the results. But this isn’t action. It’s a perfect notion of reaction where everything you are doing is a response to what you’ve already done. The essence of action is not to produce results but to bring forth [create] the possibility of what can— but does not yet—exist.

It’s important to remember that language is not just about words. Mathematics is language. Music is language. Dance and art and meaningful body movements are language. Even occupational activities such as computer programming, carpentry and automotive repair have specialized language. If you are competent in any domain and have the distinctions (the language) of that particular field, you are able to navigate in that world and coordinate with others. By competence, I mean the ability to make and keep promises.

Making and keeping promises is an example of how human beings create the future. A promise is an action in language. Specifically, a promise is a commitment and an offer you make to others to generate a particular result at a particular time in the future. Each promise only occurs in the present: you cannot make a promise yesterday and you can’t make a promise tomorrow (because when you do, it will be today again).

If you repeatedly don’t keep your promises, others will stop taking you seriously and won’t trust you to do what you say you will. If others won’t accept your promises, then coordination will be poor or non-existent. If a promise is a commitment and an action which occurs in language, this raises interesting questions.

Is artificial intelligence capable of making a promise? Is GenAI, with all of its conversational and language-generating capability, also capable of action? Can GenAI make and coordinate commitments independent of human beings?

The basis for Yuval Harari’s remarks in his talk “AI and the Future of Humanity” is the fact that ChatGPT and breakthroughs in other emerging AI applications are demonstrating their capacity to generate and process language in much the same way human beings do. Once this capability exists, he argues, machines don’t need consciousness to profoundly impact and possibly dominate human civilization in a manner analogous to the technological fall of humanity portrayed in The Matrix movies. In that movie, which was based on a philosophy found in Jean Baudrillard’s 1981 thesis “Simulacra and Simulation”, humanity lived inside a virtual reality created by an artificial intelligence. (Baudrillard had wondered how you would know you were living in an artificially constructed and controlled reality if you were living inside it. In The Matrix and its three sequels, with the exception of a few outcasts, people did not know their experience of the world wasn’t real.)

I think all of us are inclined to anthropomorphize our tools, especially technological tools.  I know I do. I laugh at myself when I raise my voice to the automated answering system that wants me to explain what I am calling about when all I want to do is speak to an actual person. My wife sometimes thanks Alexa for reminding her of what’s on her shopping list.

Machines can already manipulate responses and are increasingly capable of producing intimate conversations with human beings.

When Amazon tells you your package will be delivered on a specific day, its automated messages are making a promise on behalf of the company. If we assume that what they say will happen, we’re responding in exactly the same way we might if a person had made the promise. If we get upset should Amazon not let us know the package will not arrive on time, our response is also similar or the same as when a person fails to fulfill a commitment. The result is inevitably the same: we stop believing the promises and don’t trust the system or whatever entity is involved. If the AI equivalent of a ‘customer service’ representative tries to repair the damage to that trust, the same will probably be true. If it gives excuses, they will rarely make any difference. If it takes responsibility, acknowledges a mistake and makes a new promise, we might let it go and move on.

Action/commitment in language is constantly happening in most of our conversations. It happens in both our speaking and our listening. For human beings to create anything, there must be coordination. And that coordination cannot happen unless you as the listener are also engaged in a committed way and connecting whatever actions/commitments you are making to the actions/commitments being made to you.

Therefore, if you and I are listening to a GenAI promise as a promise and adjusting our own commitments accordingly, then GenAI is committing and coordinating with us. When we are unable to distinguish something written or spoken by GenAI and something generated by human beings—and we’re already very close—I see there is no difference. On the playing field of life, it’s all action. Who the actors are is irrelevant.

The question to then ask is what can human beings bring to the conversation that a machine cannot? My answer is context, which is what I will write about in my next post.