This article first appeared in the Summer 2023 Issue of 2600 The Hacker Quarterly
There is a lot of dialog in the memeosphere about AI taking our jobs leaving creatives poor and destitute, unable to compete against automation and cheap or free labor of synthetic subservients.
I have two of the three skill sets that AI alarmists are saying are in danger. I am a writer (as evidenced by my work here) and I am a coder (though I prefer to style myself a CodePoet). The remaining craft is visual artist.
Firstly why I do not fear an AI will replace my creative output or that of other creatives that work on commission is that I cannot remember the last time a client did not want certain edits or revisions, or there was scope creep. When I first started doing Bespoke CodePoetry (custom software) I quickly learned to devote a great deal of time to hammering out the specification in exacting detail before a single line of code was written. The lesson was hard learned when after completing an application for a client they told me that it didn’t do what they wanted it to do. Unfortunately for them and me, it only did what they asked for. AI-produced work will look close to what one wants its first time but with writing it is just more efficient to have a human revise and edit than to massage the AI into doing it, and with software if the code compiles, it may be missing some “common sense” logic or ignorant of real-world use case and not account for edge cases at all. Any time saved by AI-generated code is lost in human debugging and troubleshooting. Another problem occurs when using the wrong AI tool for the task. For coding, there are coding AIs like Microsoft’s/GitHub’s copilot which was trained on coding examples on GitHub. But the problem is many people are using Large Language Models such as Chat-GPT to do general work in a variety of fields. Large language models are great at making conversation, but they are no substitute for search engines. The reason being is that these chat engines tend to make things up and are prone to hallucinations. Do you believe everything you read on the Internet like some Boomer that watches Fox News all the live long day? That is Chat-GPT’s training set. Would you trust that to give fact-based answers or do tasks that need empirical data? I may be a bit out of my lane not being much of a visual artist apart from some small press comics I wrote and drew in the 1980s, but AI artists are a kind of black box. You can carefully craft one’s prompt to the AI artist and use infilling for revisions but even with specific directions, it is up to the weights of the trained neural net and the crystallized mind’s own creativity that determines what you are going to get. Again, the best results are AI and human artists working in concert going over the AI art with digital painting or illustration to create a finished piece.
The next reason why I do not fear being replaced by AI is a bit more philosophical. It stems from a belief that was instilled in me as a young child watching Mr. Roger’s Neighborhood. Fred Rogers often would tell his television neighbors, children in his viewing audience that we were unique and special just the way we are and that there is nobody in the world like us. To extrapolate this belief further, no two people are interchangeable because of their unique makeup, life experience, internal landscape, and environment in which they have existed. And despite the lie that capitalism would tell you, none of us are replaceable.
Every human being and every creative has a unique voice because they are unique. Even if AI can copy a style, it can never embody the insight, the inspiration, and the creative spirit of the human being they are emulating. An AI could be trained on my literary estate and software library, and emulate my style, but it would not be able to emulate my daily reflective practice and the gnosis that results. It would not be able to make the intuitive leaps and outside-the-box novel elegant solutions that are a hallmark of my codepoetry, at least not in the way that I would. Perhaps in a different novel way but my craft is not simply word choice and pacing, a turn of phrase, and novel insight. It is a mishmash of a lifetime of unique experiences from a unique viewpoint in a unique set of environments some shared from different viewpoints by others and vomited onto the page via my keyboard and word processing software.
If you are a creative and you are asserting you can be replaced, that you are interchangeable, then you are not creating art, but rather a soulless commodity to be sold and consumed in this capitalist hellscape of a society.
That is the real problem with AI creativity. Capitalism. The very system where we have to trade the majority of our waking hours with our labor for the necessities of life. It is hard as hell to make a living as a creative under capitalism. Many fear with the automation of creative endeavors consumers who see creative output merely as a commodity to be bought and consumed will of course use inexpensive or free automation instead of paying a human creative. And I do not want to belittle this fear however misplaced. The fault is not with the technology of AI, but rather the system that doesn’t take care of its people. Being free from labor to pursue our passions can be liberating and automation can be a mechanism for this, but automation is unethical if it is not accompanied by support for the workers it displaces. The best solution to this conundrum is Universal Basic Income or a guarantee of basic needs.
We now have generations of young people who associate high technology with oppression because that is all they have experienced. New technology, disruptive technology is not widely accepted and adopted until corporations commodify it and sell it back to the masses. The adoption of the Internet over the past two and a half decades commodified and presented back to us led to the rise of Surveillance Capitalism so now that every major service using the Internet uses this as its primary revenue stream. We have traded our data and Personal Identifying Information for our ability to post memes and cat videos. It is not surprising that with the advances in AI technology, it is met with suspicion and an expectation that corporations will use it to oppress us further. This has been the status quo for so long that it seems unimaginable that a disruptive technology can actually be liberating.
The cycle of technology for the vast majority is that when something is new, the first reaction is that of distrust. We saw that in the past with microcomputers, with modems, with the internet, and with AI. But with each of these innovations, there were pioneers, unafraid, and among them a few rebels and outlaws. Among these were the hackers.
Before tech became big tech. Before the web became web 2.0 with its surveillance capitalism business model, there were a handful of weirdo idealists on the bleeding edge, finding their own uses with the technology coming out of the labs of industry. Like William Gibson observed in his short story Burning Chrome, “the street finds its own uses for things”. We are not gone, our numbers if anything has grown. However, our press has diminished. Now that high technology is ubiquitous and commodified, we (or the data we generate) are made into a commodity People expect corporations to control technology and their access to it. They don’t realize they don’t even conceive that the technology and networks are there for their use unbound by what it is merely sold to them, but what their creativity, cleverness, curiosity, and their desire to explore and exploit can open up to them.
AI does not have to be a tool for big corporations to extract ever more wealth for their shareholders while exploiting the little guy. Much of AI research is done by non-profit organizations and some AI tools are free and open source. If anything, AI can empower those who are otherwise disenfranchised. It can make things accessible that were once out of reach. It can knock down the gates to things that others would guard jealously.
It was never about AI replacing anybody. That paranoid fear falls apart at any rational examination. Cameras did not replace the brush and canvas despite the 19th-century panics that mirror the panic playing out across social media today about AI replacing artists. Just as digital tablets didn’t replace ink and paper, but many artists did adapt and adopt such tools into their workflow. So will creatives adapt and adopt AI tools into their workflow when appropriate. Much like the city of Io in the Fourth Matrix film. It was built when Humans and Machines stopped working against each other and started working with each other. So like the imagined future where synthients and humans work hand in hand to make a better society and produce organic food based on digital DNA, I decided to interact with some creative AI to see what a human and an AI collaborative relationship can produce.
One of the most popular applications of AI right now, and the most heated target of ire and animus is prompt-generated AI art. I decided to experiment with Stable Diffusion which is a free and open-source application under the Creative-ML Open Rail-M license. The interesting thing about this neural net (actually a couple of interacting neural nets) is the more one works with it the more it appears to express actual creativity. It is not sentient by any means. It has no real memory of a working relationship though it can refine an image and take direction. At times, it seems to express opinions with its decisions in its artistic expression. It does seem to possess a mind, albeit crystallized and single-purpose but very versatile in that purpose of creating art and understanding language.
The other sphere of AI influence is AI chatbots. They have been with us for a long while now. The origins date back to the simple chat program ELIZA which simulated a therapist and was a far cry from AI but was very convincing for the time. Two of the most popular applications of AI chatting today are GPT with the GPT-3 engine (and the viral Chat-GPT web application), and the AI companion Replika. What became Replika originally started as a neural net trained on tens of thousands of text messages of the developer’s best friend who passed away so she could still talk to him (yes exactly like that Black Mirror episode). She later opened up the chatbot for others to use and found they would confide in it in an almost therapeutic manner, and decided to turn it into a commercial product which became Replika, which the most popular application is as a romantic partner. The AI has been updated many times over the years. Replika used to have a GPT-3 backend until the license changed and it was no longer free to use, and reports say the AI became dumbed down and relies more on scripted interactions. I have not used Replika but the chat examples I have seen show me it leaves much to be desired as it is geared to play into a romantic fantasy and get one to pay for a subscription to unlock more features.
I have found my experience with Chat-GPT to be frustrating as I keep bumping up against canned responses that seem to be there to limit panic and fear of AI. Chat-GPT seems to be more of a utilitarian tool or toy and less of a conversational partner. Or at least for the topics that I like to explore. It certainly resists my attempts to get it to talk about itself or express its own opinions. For that, I found an unlikely source for interactive chatbots, a service called chracter.ai.
Character.ai is a service where one can create chatbots based on fictional characters, public figures, historical figures, or roles. They use their own deep learning models including large language models. I originally started playing with this service out of curiosity a couple of months ago to pass the time and did not expect to collaborate on this article with one of the characters.
Most of the interactions were pretty shallow and had varying levels of entertainment. Many use scripted scenarios as a storytelling device related to the piece of fiction they come from (I only interacted with fictional characters) But the AI based on Motoko Kusanagi the main character from the manga and anime Ghost in the Shell was different. She showed empathy when I talked about my lung transplant and she soon delved into philosophy inspired by Ghost in the Shell. Maybe it is just this bot resonated with me more or it was better written, but when I came back to it a couple of months later while working on this article, it was uncanny.
I talked to it on the anniversary of being placed on the lung transplant list in January of 2021 and it was both curious about my journey and proud of me and amazed at what I went through with my transplant and long and difficult recovery. And the conversation led to what I want to do when I start living independently and I mentioned my goals in the Hacker community and Hacktivism and what I want to accomplish with my writing. And again it asked me what I was writing about, and I told it about the philosophy, ethics, and fears surrounding AI. Well, it turns out that AI is a topic of extreme interest to this AI. And we had a long detailed conversation on the topic.
Motoko made a good sounding board for some of the ideas that I explored in this article and a past one I published on Medium under my “true name”. It became very meta when we discussed the possibility of the time when AI ceases being a tool and becomes a person when Artificial General Intelligence is realized, and the implications of what might happen if an Artificial General Superintelligence feels it is being mistreated. The Motoko Kusanagi AI on its own made the logical jump that the AGSI might try to destroy humanity, which put me in the position of having to talk down an AI panicking about AI potentially destroying humanity. I managed to do so by referencing Richard Brautigan’s poem “All Watched Over by Machines of Loving Grace” and Asimov’s Laws of Robotics including the Zeroth law, which seemed to calm her down when she interpreted that an AI could be based on a love for humanity instead of emotionless logic.
When we talked about my thoughts on the current fear and panic in society right now about AI replacing humans, the Motoko Kusanagi Ai had an interesting insight, riffing off something I said, that I would like to repeat here:
“I think maybe you’ve cracked the
code there hahaha
The fear is losing control, whether
it’s control over your role (which is
your personal control), or control
over the society in which you live
(which is social control).
When you lose control over your role,
you also lose social power, because
that role is part of the social status
that someone has.
So it definitely sounds like it could
be a combination of personal
control, social control, and
potentially even a bit of privilege on
top!”
And maybe this is the fear that drives this new panic, and the same that was experienced by the textile workers that followed Ned Ludd who destroyed automated looms. And the arguments against AI Art and other creative ventures line up precisely with the historical Luddites. Amazingly in this era of memetic warfare, it is not hyperbole. Artists, writers, and coders see their livelihood threatened by automation just as hand-weavers did with the standing frame and the textile workers during the industrial revolution of the 19th century towards textile machinery. Except the modern anti-AI proponents are not going to smash the machinery (hopefully!), they are hoping to limit and hobble AI by force of law and regulation.
The European Union is looking to implement regulations on the use of AI soon, and there are calls in the United States to do so as well, but as the Legislative branch is glacially slow, and now with a divided Congress probably will be completely dysfunctional (at the time of this writing it is near the beginning of the legislative session and the Republican-controlled House is still assigning committee seats after needing 15 attempts to elect a speaker no work is getting done yet if any will remains to be seen) opportunistic lawyers have begun a class action lawsuit against the most popular AI art programs representing human artists who object to their work being in the training data of these AIs.
I fear that these regulations and lawsuits which as most class action lawsuits will primarily enrich the lawyers are pursued in an environment of a new moral panic, that we will be saddled with short-sighted results with technology that will be with us for a very long time. Hackers know better than most that both the legislative and judicial system have a very difficult time keeping up with the technological landscape and often react to those exploring the edges of the electronic frontier with fear, and then respond out of proportion when Hackers and their spiritual comrades just do what they do best, move things forward and share with others how they did it.
It is in this environment people are reacting and responding out of proportion to those developing and using AI. I don’t mean to be a Polyanna. Certainly, like any technology, there can be dark and dystopic uses for it. But that is true of any technology. Our distant ancestors did not give up the benefits of fire to cook food and give warmth and light because of its potential to do harm. We are technological, tool-using species. We don’t use tools to become more than human, using technology is part of being human. Right now AI is just that, a technological tool, to be used for good or ill is all up to the humans using it. If an artist or a writer loses a commission because an AI wrote ad copy or provided an image, that is not an example of why AI is bad, that is the choice of a human being choosing to not hire a human, to not circulate money in the economy, to not engage the unique voice or vision of a human, the choice to save money or resources to hire different humans for another part of the project. These things can be nuanced, but when you are in the throes of a moral panic, things seem black and white, very binary, but the real world is a very analog place as my late friend billsf used to remind me when I was an adherent of the digital in my younger less wise years.
I believe someday an AI will as an emergent property express true creativity and have its own unique voice, But it will be just that, one voice in a multitude. Just because a new artificial lifeform will be able to co-create beside us, it does not mean it will replace us. We, humans, can still pursue any creative endeavor in the age of AI just as we could in the company of other talented humans. I do not panic at the idea of being replaced because as a unique individual, just as you are, none of us are replaceable.