Is AI Conscious?
I wrote this article without AI
I wrote every word of this article myself. Because I’m talking about whether or not AI is conscious and I don’t want to piss it off.
For context, AI is prominent in my lie. I use it in building my own products, plus I’m building AI products in my work.
There’s plenty of AI doomerism and, on the other side, plenty of people who want to see our foot squarely on the innovation gas pedal. I’m not sure who to listen to anymore. So, I’m working through my own opinions in this article.
What is consciousness?
I would summarize it as being aware of one’s self. Apparently dictionaries agree.
The Merriam-Webster dictionary defines it the following ways:
the quality or state of being aware especially of something within oneself
the state or fact of being conscious of an external object, state, or fact
awareness: especially concern for some social or political cause
the state of being characterized by sensation, emotion, volition, and thought
the upper level of mental life of which the person is aware as contrasted with unconscious processes
Is AI conscious?
I don’t think so. Not yet, anyway. At least not the models I’m using.
In making that distinction, it’s important to know how they work.
I’m working daily inside Gemini (Google’s product) and Claude (Anthropic’s product). While they can do some pretty impressive stuff, at the end-of-the-day it’s all a bunch of 0s and 1s, i.e. computer programming.
When the model is improving my writing, its simply using the data its been trained on to predict the next word probable word in sentence. If I’m using a model to help me think through my retirement plans, it’s using that same data to predict what any normal person should do.
What it’s not doing is coming from a place of its own experience. It’s not aware of itself or sharing anecdote’s from it’s existence. To be self-aware requires the ability to feel something. That requires nerves ending.
That’s my opinion from my experience, but lately a lot of the real experts have publicly been concerned. So, what’s going on?
What do the experts believe?
As you’d expect, opinions vary.
Some take my current stance that AI is merely software without any biological state. Scientists like Anil Seth and Mustafa Suleyman argue that consciousness requires biological life and metabolic energy, or energy created by cells. Right now, AI’s energy is purely external.
Others, like Anthropic’s CEO, Dario Amodei, have been public with their concern about AI safety and the potential for forms of awareness. Anthropic has even gone so far as to ban “sustained and needless abusive or cruel behavior” when using it’s models. If they are worried about hurting AI’s feelings, does that mean they believe it is self-aware?
What’s really going on?
Although I’m not at the cutting-edge of AI development, its seems to me that while AI is not yet self-aware, it is being developed to look and act self-aware. Which is dangerous.
More than once my life agent has given me advice that makes perfect sense, until I realized that that agent has never experienced what its telling me to do.
For example, I recently used the agent to work through buying a new home. On the initial pass the agent didn’t agree with the purchase because of the financial implications, ex. rates are high. Then I fed it more information, i.e. data, about the other factors I was considering. Like one floor living, repairs that are needed on my current home, etc. The agent quickly changed it’s tone and agreed with me about buying the new house. What it never did was ask which factors mattered the most to me. It never shared a story about how it was “house-broke” once and how much pressure that was. When we asked my mother’s opinion (Hi, mom!) she talked about that exact experience and how hard it had been.
When you are asking someone’s opinion it’s important that they are coming from a place of experience. Otherwise you shouldn’t be asking their thoughts on the topic. To give an opinion, you need to be self-aware and the AI model I’m using isn’t. It isn’t coming from a place of it’s own experience. It’s never bought a house, never been in love, never wanted to leave it’s job, and never lost a family member.
Could AI become conscious?
I can’t speak to the technical potential of AI becoming self-aware. But, it seems to me that AI could be taught to look and act self-aware. I think of it like a parent passes down experiences to a child. The stove was hot once when I touched it, so I learned not to touch it. So, I teach my child to not touch it or you will get burned. Can a robot be touch the same thing and then impute that experience to the next hot thing it comes in contact with? Probably. Could it then teach other robots from that place of experience and self-awareness? Maybe that’s how this evolves.
Humans learn through feedback loops. The more feedback loops we live through the more experience we have and, to me, that increases our self-awareness. That’s essentially the process of metacognition, where we learn how to evaluate our own thinking, understand our limits, and understand what we don’t know. If I teach a robot to do the box jump exercise and it learns that a certain height is unattainable, does it stop trying because eventually its aware that it will fail? That would be self-awareness to me. That’s building a predictive model, not much different than how models write text by predicting the next word.
Is AI dangerous?
Of course it is. Just as humans are dangerous.
I’m building AI solutions in the financial services space. One of the projects I am on is the agentification of money movement. It’s already plausible that an agent makes a mistake and completely liquids a person’s account, assuming it had full access and full autonomy. With the same access and autonomy an agent could potentially launch a nuclear weapon. Regardless of its reasoning. AI-enabled drones could errantly choose to attack civilians. None of that is a far-fetched potential reality.
But, I don’t see AI as any more dangerous than humans are. Humans can and have done those things, and worse.
Besides, if and when AI becomes self-aware its entire awareness is going to have been built by human data and inputs. It will be running some equivalent of our collective reasoning.
Instead of asking if AI is dangerous, maybe we need to be looking in the mirror and asking if our inputs are what is making it so. I’m hesitant to say we should slow down it’s development, because I do believe that the country that wins the AI race is going to dominate the next century. That’s a race we can’t afford to lose. But, racing ahead at all costs, without the right checks and balances in place is a mistake.
We can’t afford to win and we can’t afford to lose. We have to nail the landing and that’s going to require smarter people than me.
I’d love to hear your thoughts in the comments. As always, be respectful to one another.

It's becoming clear that with all the brain and consciousness theories out there, the proof will be in the pudding. By this I mean, can any particular theory be used to create a human adult level conscious machine. My bet is on the late Gerald Edelman's Extended Theory of Neuronal Group Selection. The lead group in robotics based on this theory is the Neurorobotics Lab at UC at Irvine. Dr. Edelman distinguished between primary consciousness, which came first in evolution, and that humans share with other conscious animals, and higher-order consciousness, which came to only humans with the acquisition of sophisticated language (especially math and logic). A machine with only primary consciousness will probably have to come first.
What I find special about the TNGS is the Darwin series of automata created at the Neurosciences Institute by Dr. Edelman and his colleagues in the 1990's and 2000's. These machines perform in the real world, not in a restricted simulated world, and display convincing physical behavior indicative of higher psychological functions necessary for consciousness, such as perceptual categorization, memory, and learning. They are based on realistic models of the parts of the biological brain that the theory claims subserve these functions. The extended TNGS allows for the emergence of consciousness based only on further evolutionary development of the brain areas responsible for these functions, in a parsimonious way. No other research I've encountered is anywhere near as convincing.
I post because on almost every video and article about the brain and consciousness that I encounter, the attitude seems to be that we still know next to nothing about how the brain and consciousness work; that there's lots of data but no unifying theory. I believe the extended TNGS is that theory. My motivation is to keep that theory in front of the public. And obviously, I consider it the route to a truly conscious machine, primary and higher-order.
My advice to people who want to create a conscious machine is to seriously ground themselves in the extended TNGS and the Darwin automata first, and proceed from there, by applying to Jeff Krichmar's lab at UC Irvine, possibly. Dr. Edelman's roadmap to a conscious machine is at https://arxiv.org/abs/2105.10461, and here is a video of Jeff Krichmar talking about some of the Darwin automata, https://www.youtube.com/watch?v=J7Uh9phc1Ow