Conscious machines. Software that empathizes with you. Code that can think…
Have we all gone mad?
Let’s get the virtue signalling out of the way, before I continue. I am not a technophobe. As my daughter can testify to you, I once sang to Shazam because I believed in its endless power to identify tunes. It didn’t work. Still, Shazam is one of my all-time favourites. Also, I typed up my master’s thesis on an electric typewriter, handed down by my ever curious grandfather. Thus, there is hardly anyone more grateful than me for the advent of the word processor. And with my daughter thousands of miles away from me, video calls are an absolute lifeline for my heart and soul.
Having said all that, I wonder though: How is it that human beings can seriously claim or believe the delusional concept of “conscious machines”? Because marketers call processing “thinking”? Because the machine output is getting more linguistically fluent? Whenever I hear the qualifier “fluent” in connection with language proficiency, my first question is: “Does it only have to be fluent, or must it also make sense?” Fluency is easy. You just ramble on and make it sound important. It is no yardstick for mastery of anything, though.
The same is true for machines. The output of machine translation, just to take an obvious example, has indeed become more fluent. It also has become more accurate to a certain degree and in certain settings. It does have its use for anything inconsequential. Still, it remains a guessing game of probabilities. The underlying Large Language Models are basically playing dice with words. Interestingly, these systems get more dangerous the more fluent they get, because fluency suggest a level of trustworthiness that is simply not justified. A bit like with pompous, motor-mouth speakers.
Professionals do not play dice. They know what good looks like.
As Einstein said: “God does not play dice.” Neither do professionals. Real pros in whatever profession make informed, conscious choices. They know what good or excellent looks like and work towards it. They gather insights to feed their three decision making instances – the brain, the heart and the gut. They sense the vibes around the matters at hand and how things resonate with. They ask questions to make sure they get the full picture. Having done all the gathering, they digest the intake, applying their knowledge, experience, critical thinking skills, and their hopefully well-trained gut feeling, while all the time being mindful of their own filters and triggers. Once the digesting is done – which can take mere minutes or several days – the result of that process is fed back into the world as best as one can and the environment allows. When feedback comes back, it is basically rinse and repeat.
This is roughly how conscious decisions are made, leaving aside the funky possibilities opening up once you access meta-cognition or a completely different source of knowing.
Machines don’t care
No matter how all of this beauty plays out in individuals, one thing is for sure: It requires a human body and human emotions. Machines have neither. Thus, they will also not fret about potential outcomes, no matter how much they are potty-trained to give the illusion of being considerate. Real fretting is visceral. It is result of real caring, even if it’s just caring about covering your proverbial. Machines don’t care.
You cannot hold this against them, because they are just not wired that way. Also, machines actually don’t have to care, because they never have to answer to anyone for their so-called decisions. Your bot body gives you a wrong translation that creates misunderstanding and actual damage down the line? Oops, sorry. It gives you wrong “medical advice” and you suffer the consequences? Buddy, you know you should not have relied on me, man. I hope you get better soon.
A very interesting word that is used to describe the machine botching it up is “hallucination”. That’s cute. Hallucination in human beings is about sensory perceptions. A machine has no senses. Consequently, a machine is not hallucinating, it produces bullshit. It produces falsehoods. If a human being were to do this, she would have to face the consequences. She would be personally liable. Machines are no liable. But you will be, dear reader, if you sign off on AI info and decisions that you do not entirely understand, because your senior management “encourages you” to use the machine like this.
Now, it is also true that currently not every human being on this earth acts as consciously as described above. Perhaps their internal decision making instances are not properly in sync yet, or the environment they operate in does not give them the time or space to make conscious decisions. This does not mean, however, that they can’t. They can. They just got to learn. If we all got better at this, this would indeed be major progress in the grand scheme of human evolution.
Why would anyone want machines to outsmart human beings?
We might never know, though, how much real progress we can make, if we allow human beings to be gradually de-skilled through machine use and to lose trust in their own abilities.
There is a plethora of studies on the depletion of cognitive abilities of human beings through the use of digital entertainers and helpers. Our kids can concentrate less and less. They lose the ability to read for meaning. This is not progress.
Many years ago, I sat with one of the finest minds in computer science in Germany – an academic as as well as a practitioner – to do research on algorithms for something I was writing. He sat with me for three hours, and it was probably one of the most interesting and enlightening conversations I have ever had. He found fault with the “I” in AI, and he said to me verbatim: “Intelligent ist nur der Mensch” – only human beings are intelligent.
Perhaps we should thus take a step back from all the AI hysteria (which, by the way, shares all the characteristics of the subprime hysteria of 2008/2009) and ask ourselves some very basic questions, such as: Why, the hell, would anyone want machines to outsmart human beings? What type of aim is that?
I totally understand why we would want performant tech that serves us. Great tools in the hands of discerning experts that help make further progress in medicine or engineering? Easy access to precious knowledge for all of us? I am all in on that. However, tech that keeps us from expanding our consciousness because it erodes our cognitive abilities? Tech that makes us lonely, depressed and dumb? Tech that is used to harvest skills and knowledge from human beings to train the machine and then ultimately replace human beings? Honestly?
If you use the machine to do away with entry level jobs for graduates, for example, where do our future lawyers, engineers, doctors, translators, nurses and about everybody else gain the experience which will help them to build their knowledge and skills, their confidence, their priceless gut feeling, and, ultimately, their ability to take conscious and responsible decisions? Where will be their spaces to fuck things up and learn? Who will show them the ropes, and in which language?
Standardized minds don’t innovate
Human language, the ability to have an internal dialogue, express ourselves and cooperate with others, is one of the main things, if not the main ability, that make us human. The language that is pushed through LLMs, though, makes language its whore. It is a soulless attempt to press the beautiful chaos of applied human language into mathematical models by ripping it apart into tokens and glueing it all back together on the basis of probabilities and English as a master. The more we will be subject to this standardised and sterilized language, the more standardized our thoughts will become. The more standardized our thoughts are, the less true novelty and innovation will come into the world.
Performing a life instead of living it
AI, in particular in combination with so-called social media, gives people endless opportunity to perform a life without actually living it. Such as giving the impression of being a writer without putting up the effort of having a painstaking internal dialogue searching for the words, the sentence structure and the rhythm that gives the best possible expression to what you carry inside.
Without being able to have an effective internal dialogue, there will be no individuation in the sense of Carl Jung, i.e. becoming a whole human being. The more this ability erodes, the less we will be able to have meaningful and nuanced conversations with our fellow human beings, which is the prerequisite for making any social contract work. We have seen increasing polarization in the world. It is bound to get worse the more we lose nuance, because with nuance we lose discernment, without discernment we lose the ability to deal with other realities, appropriately relate to our fellow human beings and cooperate effectively.
Some people say AI will be all hunky dory, and we will eventually all live fulfilling lives in leisurely abundance. I actually do believe that this is possible. For this to happen, though, the way of shaping and using AI will have to change radically. It cannot be our master. It must be our servant. It must be radically “human beings first”.
The essence of totalitarianism: Making human beings superfluous
I have my doubts as to whether the current usual suspects will get us there. Looking how they make people redundant to pay for compute (or push the stock price) or how they speak about human beings, strongly suggests that they are not too fond of humanity and don’t think too highly of it.
Several years back, the former Google CEO Eric Schmidt gave a speech at the technical university of Munich (TMU). Amongst other things, he said that employees who just come in to do a 9-5 job (i.e. the people who keep the show running on a daily basis and probably take care of their families and communities when they get home) are “not very interesting”. In his biography of Elon Musk, Walter Isaacson describes how Larry Page, another Google man, accused Musk of being a “specist” when Musk spoke about his love for humanity and human consciousness.
Standard Chartered CEO Bill Winters just recently told investors that the next job cut of 7000 positions affected “lower-value-human capital”. Ladies and Gentlemen, take it from a German, that’s pure totalitarian playbook speak. It is testimony to the assumption that you are a superior being who is entitled to decide what other human beings are worth and what to do with them. He apologized later after some uproar, but he gave us a very juicy look into how his mind works and how he obviously thinks he must speak to investors to sound all sexy.
In The Origins of Totalitarianism, the philosopher Hannah Arendt argued that the ultimate goal of totalitarian regimes is to organize society so that individual human beings become completely superfluous.
Don’t panic!
So, what the hell do we do? Don’t panic, people! Much of it is just a big circus, and many of the clowns will eventually leave the stage.
In the meantime: Ask questions. Question yourself and those around you. Use your own mind. Do not only listen to what people say but how they say it, and challenge the pompous. Speak up. Be mindful when you use anything digital. Don’t get paranoid, but be mindful. Ok, perhaps a little bit of paranoia might be warranted here and there.
Most importantly though – please go out and mingle. Know that you are unique and that everyone you meet is unique as well. That there is more awesomeness inside of you than could ever fit on any chip. Let your consciousness flourish and expand – go for walks, meditate, read good books or just sit and stare at the ocean/mountains/hills/fire. Stretch your cognitive skills and trust them. Get talking, exchange your unique stories and takes, and combine them into something great. Laugh, fight, make-up, have great sex or not, travel if you can, share good food, or just sit next to each other in silence. Dare to love and trust.
And, by all means, take your socially-challenged tech bros along, so that they might lose some of their fear of human interactions and think a bit less about how to control humanity by standardizing complex human beings into neat boxes that they can handle. While you are at it, don’t forget to tell them that you love them.
