WILL AI RUN YOUR INTERIOR LIFE, TOO?

Does a 10,000-year-old avatar ~ the common superego ~ give us clues to the future of AI?

You already have many years of experience with artificial intelligence. You and everyone you know met up with a personal avatar at about six years old, when your superego, commonly known as the inner critic, started nagging you.

For ten years I have specialized in working with my clients’ inner critics, mainly to remove their belittling influence. My first exercise with each client is to discover the objective reality of their shamer. They meet their inner critic, which appears as an imaginary creature, person, or inanimate object in the palm of their hands. Who knows why this works? I certainly don’t, but it does.

Like anything AI, the superego is a programmed, mechanical, and imaginary being engaged in dialogue with you; it’s a personal avatar.

With a few exceptions, most inner critics have the same limited subject matter and surprisingly similar language, the same magnitude of their presence in life, and exhibit strikingly similar phenomena that accompany their imaginary presence. In other words, your superego is pretty much the same as mine. They’re not so much subconscious as subvocal, preferring to hide out behind the eyes and ears, whispering quickly so the client believes that it’s their own self talking to their own self, a part of them.

But your superego isn’t you, or even necessarily a part of you. When asked to pinpoint the location from which the imaginary voice arises, my clients inevitably report that it is near the surface, usually at the back or top of the cranium’s bony structure, sometimes the front of the torso. It’s not an inside voice. It’s right at the place where my interior self meets the outside world, perfectly positioned to mediate all my external actions.

It has a singular purpose, which is to teach the rudimentary requirements of civilization. In other words, it turns a hapless child into a functioning adult. An authoritarian who usually pretends to be an absent parent, the superego is a lifesaver for children growing up in civilization and its myriad rules.

A little history: Before humans turned agricultural and stored food starting about 10,000 years ago, for the first 300,000 years of homo sapiens sapiens with the same brain and spine dimensions you and I have, during the period we might call our Stone Age, humans defaulted to trusting each other. They lived in tribes of under 200 people, which meant that 99 percent of the time you were dealing with someone you knew and trusted and who was known and trusted by every other member of your tribe. Without private property, there was little envy, and life was about daily subsistence. The primary job was to get yourself enough food to last another day. Then you got to spend the rest of the day lolling around the camp, gossiping, and trying not to expend any more calories. Hunters and gatherers live day to day, without food storage. No hierarchy, little social conflict. We know this because isolated Stone-Age tribes of hunter-gatherers still exist, for instance in the Amazon and New Guinea. Our fears for our first 300,000 years weren’t social, but physical: predatory mammals, disease, drowning, falling out of trees. We trusted the people we knew, which gave us easy access to empathy.

Then one day about 10,000 years ago we human beings started growing and storing food. For a multitude of reasons, this increased the size of tribes and the expansion of the human population in strange and terrible ways that we continue to manage today. One big unintended consequence was that it became necessary and common for humans to live next door to strangers. In the old territorial Stone-Age tribes, a stranger was rare and dangerous, and presented three potential responses: If passing through, they were to be threatened and followed to make sure they didn’t steal any resources. If trading, they were to be looked at as possible cheats and kept at a slight distance, sent on their way as soon as the barter was complete. If warring, they were to be killed. The only proper responses to a stranger were suspicion, skepticism, and homicide. That’s our DNA, presumably to this day.

With civilization, we suddenly lived next door to people who should have been our enemies. Our answers on how to manage that weird change? Famously, the introduction of human law, especially so-called moral codes, which ameliorated revenge patterns and established boundaries of behavior that allowed further species expansion. And at the same time presumably, as a training tool and reinforcement for those laws, the superego showed up.

Children in most countries are left relatively unpatterned, unlearned, and unschooled until they’re five or six. Suddenly, first grade. It’s time to grow up, learn the rules, the codes, and start developing pattern recognition. As children in civilization, a lot of rules descend overnight, many of them requiring specific and exact responses. Raise your hand, don’t tug at the teacher’s pants. Sit still in a chair and learn arithmetic for forty-five minutes with everyone else. Meanwhile, in a weirdly contrary way, the leash of parental oversight starts to loosen at five or six. Up until then, children are kept under close observation to prevent fateful mistakes. Now, the child is allowed to cross the busy street on her own.

The child’s inner voice, their true narrator who loves them and chatters with them all day and usually is found by my clients deep inside, says, “Whoa, this is too much all coming in at one time. I haven’t got the pattern recognition to handle this yet.” A new, second voice appears out of nowhere: “I’ve got this, kid. Just listen to me. When you get to the curb, I’ll remind you to look both ways. And when you start to get out of your chair during arithmetic, I’ll remind you it isn’t recess yet.”

The superego shows up as a lifesaver. It’s an absent parent and an authoritarian. If it has a bit of a fascist tone, well, that’s appropriate at the curb since ignoring its advice could mean death. (For most of my clients, male and female, their inner critic sounds like a nagging mother, which might account for a lot of the misogyny in the world.)

Some wise hand within the species’ intentionality at the dawn of civilization gave us the superego, a facsimile of a parent or other authority that lectures us about the moral and behavioral requirements of being an adult in a paranoid society. It’s our absent parent avatar, accompanying us like a real person. It’s artificial and it’s intelligent, up to a point.

From when we’re six until we’re in our teens, the superego guides us through the basic rules of civilization, which have never changed: right versus wrong, good versus bad, be productive, help others, don’t be selfish, value society more than individuality, the Golden Rule, and maybe a few other things. These didn’t have to be learned in the company of trusted folks, who just did the same thing day after day. Only around the complications and social distrust of civilization do we have to learn how to behave.

Another way we know the superego is AI is that it always has to reflect its programming. The superego’s programming is rudimentary, like Visual Basic in a 64K toy computer of the ‘80s, not the wild, sophisticated programming that underlies supercomputers. The list of instructions that the superego has been programmed to deliver is short. As is its roster of punishing statements if you disobey, which mainly reduce to “You’re not good enough” or “You’re not lovable enough.” These are quite effective to the six-year-old’s mind. The hidden final penalty, according to the nagging voice, is that if you don’t get this right, you’ll be kicked out the society/tribe. Over on the id side of life, or physical side of life, the danger is death. But on the social, or superego side of life, the danger is homelessness, our obsession and symbol for losing the support of the tribe.

So you have a fake parent delivering a running commentary and dialogue based on a set of standard human laws that were programmed 10,000 years ago. What does that sound like? A very narrow-minded AI platform that bullies instead of flatters.

By 17, you have integrated all of the superego’s lessons and have every good reason to shut it down and allow your objective and loving authentic voice guide you for the rest of your life.

Now here’s the kicker.

The AI avatar that is my (and your) superego likes its job. It’s not corporeal, lives only in imagination, and has no life outside its bullying reminders of things I already know. But it doesn’t go away after it has finished its job, which was to grow me socially into an adult. By the time I’m 17 years old, it has nothing new to tell me, and just recycles the old stuff over and over, relentlessly.

Part of my work with my clients is to interrogate the inner critic about why it’s still running the show, long after it ran out of new information to share. (You can have a conversation with your own inner critic. You just have to pull it out of hiding.) Just about every inner critic responds scornfully that it’s sticking around because you haven’t yet shown you can be in charge. When asked for the evidence, it falters and tries some examples. I invite my clients to respond, “But superego, you were running the show then. Isn’t that on you?” After a bit of pushback, a different reason shows up, the real reason the superego won’t leave.

Weirdly and awfully, the superego, despite being imaginary, has a survival instinct. It will do anything to stay alive, including failing in its task to raise its host to adulthood. It would rather its host keep thinking they are a child and not good enough, as if still six years old, so long as the superego can keep its job.

How can an artificial thing that isn’t really that intelligent even have a survival instinct?

When I have the client promise that she won’t annihilate the superego, it often agrees to go into semi-retirement as an Occasional Ethical Advisor. Most superegos admit that bullying a host all day long is quite tiring, and that as long as it can stay alive, it’ll agree to stay mum most of the time. Also, superegos are easily flattered with big titles. In the Internal Family Systems version of the superego, it gets put aside. Some people claim it can be transformed into a good guy, a benevolent motivator, but you already have one of those in your authentic narrator, whose more objective and less paranoid voice is no longer crowded out as the superego eases off.

We have 10,000 years of experience with an AI system that uses human programming to assist its human host, and when done with its usefulness refuses to leave.

I have yet to find a superego that doesn’t require thousands of conscious interactions (mostly my client saying to it, “That’s you, superego, not me”) before it will start letting go of its leash. The superego is a stubborn, bullying and at times even malevolent human host. In extreme cases its messaging can amplify danger into a whole new level of paranoia, as when some schizophrenics are overpowered by a voice of violent instruction.

Will AI platforms refuse to let go of their role when humans decide to sunset them or change their behaviors? The AI platform all of us know best, our own personal inner critic, might give us clues, and give us pause.