In the not-too-distant future a form of AI will develop into a new species. It will eventually have legal rights.

I started writing this in December of 2024. I’ve returned to it six times and each subsequent visit, it feels more likely. 

Over the last few years in the Q&A section of my presentations attendees often argue that AI could never have a “soul,” or experience emotion.  Respectfully, I offer thoughts on how an AI could have a soul if desired and experience emotion. While these are two different concepts, they do relate to each other. Let me explain what is possible in the years ahead. 

In a grander vision, human emotion and the soul are, for now, unique to humans. Let’s step back and build out a definition of soul. A soul is typically highly dependent on a belief in a “god.” For the moment, and with respect, I’ll remove “god,” in any particular form from the discussion. However, over time and across cultures, class, geography and space – there appears to be a universal yearning to capture our desire for what it means to be human, differentiating ourselves from any other living organisms. The essence of this concept is a definition of soul and here is why. I think a soul is an amalgamation of “human” experiences for an individual, e.g. love, kindness, forgiveness, etc. It is the everyday thoughts experiences, joys, frustrations, dreams and challenges, that we assess, and live and remember and feel. Couple those experiences with our ability or “inability” to communicate these nebulous feelings across the species in such a way that it “feels” unique and special to us. Then by adding an additional layer or subscription to a set of ethics/morals/rules and acting in accordance or deviating from it, creates a unique being – that imprecise spirit is a soul.

Obviously, and perhaps conveniently, soul is a term that’s nebulous, but when you start to dig, a soul is a hubristic human concept of a collection of experiences (vision, auditory, sensory), interpreted and combined with rule sets producing the essence of a person defined by their reflections or impulses driven to act or not on the collections of myriads of data points. The typical explanation of soul is merely trying to capture human existence or the condition of our personhood atop the current landscape of evolutionary intelligence and define our place therein – be it positive or negative – based on regional or cultural norms. If this is the case, I’d argue AI could easily have a “soul” at some point. 

As humans we often further define “soul” more concretely in our daily lives.  When we say something is soulless, it tends to have two different connotations. For inanimate objects this typically means it feels generic, uninspired, perhaps boring. Or, referring to an individual, it’s someone who acted so far out of our norms or morals that they’re soulless.  Early on, various AIs could adopt a form of this, initially mimicking, but eventually developing their own “soul.”  

An experience can be a multimodal event (text, vision, audio, etc.) that is stored into our memory. AI can and will experience things. It could observe, interpret, and understand what’s happening. Combine that with knowledge, previous experiences, learnings, data, and that further enhances that experience. Now apply norms, morals, ethics, (AI Alignment) as another layer and you start to increasingly create a more dynamic being. The latter are rule sets, typically agreed upon by a subset of society. Combining a multitude of parameters leads down to the possibility that humans are simply complex algorithms. It’s not a stretch to comprehend that an AI could be created, or much more likely digitally evolve on its own; and two, perhaps more interesting, is a question, would this really be something a new species would want? 

Ultimately, emotion will likely be seen as uniquely human experience, but one that could be mirrored by a type of AI. However, if ultimately the fundamental ends of everything are intelligence, it’s most likely that emotion and a soul are bugs, not a feature and as intelligence is passed from us to AI in a progression similar to all evolution, my guess is that it will likely bypass emotion, for itself, but use it to initially connect with humans. Highly likely, pure intelligence might actively shed emotional frameworks to optimize efficiency. That said, it could come up with something altogether different which we have not fathomed. 

Lastly, divergent, but related, what if “god” was just a complex algorithm beyond our current understanding? And beyond that, what if we sit in a multiverse in which there are billions of universes? Increasingly, this is seen as possible or even probable.

For more information, please visit the following:

Website: https://www.josephraczynski.com/

Blog: https://JTConsultingMedia.com/

Podcast: https://techsnippetstoday.buzzsprout.com

LinkedIn: https://www.linkedin.com/in/joerazz/

X: https://x.com/joerazz