Series: AI Page 32 of 43

Source: The Guardian, 25 Sept. 2024:

“AI is not a tool. AI is beyond a tool,” he said. “Literally, in human history, we have never had intelligence as a technology.” […]

“I don’t believe machines should be the only creators. It’s a horrible future if you just let the machines do creative work,” he said.

Anadol said he believed it was important for artists to build their own artificial intelligence tools, and that simply using a tool someone else had built was not enough. “I collect my own data, train my own model,” he said. “I am literally co-creating with the machine in every single step.”

Here’s an archived version of the reply from NYT Magazine ethicist column on this topic, since I too can’t access the full paywalled version. I’ll just skip to the end:

As forms of artificial intelligence grow increasingly widespread, we need to get used to so-called ‘‘centaur’’ models — collaborations between human and machine cognition. When you sit through the credits of a Pixar movie, you’ll see the names of hundreds of humans involved in the imagery you’ve been immersed in; they work with hugely sophisticated digital systems, coding and coaxing and curating. Their judgment matters. The same might be true, on a smaller scale, of the fellow who sold you this digital file for a nominal fee. Maybe he had noodled around with an assortment of detailed prompts, generated lots of different images and then variants of those images and, after careful appraisal, selected the one that was most like what he was hoping for. Should his effort and expertise count for nothing? Plenty of people, I know, view A.I. systems as simply parasitic on human creativity and deny that they can be in the service of it. I’m suggesting that there’s something wrong with this picture.

My recent coverage in Business Insider got a secondary mention in this post about Sam Altman (paywalled, sorry – I use a blocker) who went on a podcast recently to reassure writers that AI will not replace them. Some choice quotes:

“There’s like a lot of bad AI writing plastered over the internet, and there’s like a lot of like bad student assignments that have probably been written by AI,” he said. “But I don’t think anyone’s serious.”

I will just say in response to that: I am totally serious.

“It’s definitely not like gonna replace coming up with the ideas anytime soon,” he said. “It’s an incredible tool for writers, but definitely not a writer.”

Well, I definitely use it to help come up with ideas. The problem is that these systems are becoming increasingly sanitized and neutered, making it oftentimes too hard to use them to really explore and cast about in the wild way that you sometimes need to do when you’re brainstorming.

While Altman thinks AI chatbots can work similarly to a “collaborator” that can undertake subtasks, he said that it would take “full super-intelligence,” or AI that reaches or exceeds human intelligence, before AI could threaten to replace human writing.

I’d like to think this is true, but after playing with the Google NotebookLM podcast generator for a few minutes based on documents I uploaded, it’s pretty clear that it will indeed and already is absolutely replacing certain kinds of writing – whether that’s podcasting, marketing copy, or other kinds of low-hanging fruit. Personally, I’m not really bent out of shape about it, because I’m a “centaur” (or a cyborg, or whatever).

“When I finish a great book, the first thing I do is like I want to know about the writer,” Altman said. “I want to know their life story. And I don’t think I’ll ever have that feeling to AI writing.”

The billionaire said that one of the main takeaways from an impactful book is the intrinsically human connection created between the reader and the author, one that he feels AI cannot replicate.

“You feel like you have this important shared human experience,” he said. “And that is some significant percentage of enjoyment of a great book to me, and I bet we keep doing that.”

I think we might be too narrowly defining still what constitutes an AI-assisted book. Seems to be here cast as something like a one-click option that shoots out a finished book. But my experience of writing, illustrating, producing, art directing, and editing over 120 short books written collaboratively between a human and many different AI systems is really different. I believe the books still deeply demonstrate that kind of human connection – and at times they do it by highlighting and holding a mirror up to the alienness of the AI-generated elements.

MSN has a thriving comment section on a free non-paywalled version of the Business Insider coverage about my work and the backlash from the end of last week. If you haven’t been able to read it previously, you can skim it over there, and the conversation developing. Some 90 comments so far over there.

It’s mostly the usual grab bag litany of complaints I’ve gotten accustomed to turning out over the course of the past 18 months or so of having people flip out on me for merely following my creative impulses and experiment with AI tools openly. So I won’t rehash any of the common negative elements cropping up over there this time.

Instead I’ll highlight some of the quasi-supportive comments below:

“I won’t use new-fangled technologies ‘cuz I’m afraid of change. I won’t fly in a plane ‘cuz if God wanted man to fly…” yadda, yadda, yadda.
A.I. is a tool, nothing more. What is written AND THEN HUMAN EDITED, is still a human creation.

It’s the spammers who throw out crap-articles without even spell-checking, let alone using any editing skills, that are the problem, NOT A.I.

It’s a tool, people, that’s all. It’s up to US HUMANS to use discernment instead of knee-jerk reactions.

If you really feel that way, you should stop using technology.

Another:

Technology only ever progresses one direction,  you either learn to adapt to it, or get replaced by it.  This Author knows whats what, and all those fighting it well, their going to go the same way the people fighting the industrial or any other technological revolution.

Okay, this one isn’t “positive” but I think it’s illustrative in how wrong it is relative to my experience of being an artist and author using these tools and integrating them into creative workflows:

As a published author, I despise the use of AI to create any content that is commercially released to the general public.  The use of AI sanitizes the creative process; it eliminates the risks, the challenges, and the rewards of something that is a uniquely human experience. The end result is just soulless content that fails to live up to what the end consumer truly deserves.

I would encourage anyone who doesn’t think my work “takes risks” to first examine the context: having thousands of people routinely tell you how you suck, and then tell me that in itself is not “risky” as a path to proceed down, let alone persist in. Then, if anyone actually read any of the books instead of trying to burn them, they might find that the actual context is incredibly risky, and incredibly *not* sanitized, and is in some instances highly challenging (I’m thinking especially of the VOMISA books, for one, but many others too). Anyway, I know I’ll never reach those people, but it feels at least good to process through these responses where they come up and understand where they score points and where they clearly fall short.

This one is a little funny:

“I made billions by not paying living, human, American employees for their labor, thus perpetuating poverty, starvation, homelessness, and crime.”
“My God, Jesus must love you because of your wealth. Prosperity has found you smart, business tycoon! Someone should do something with all these poors, btw.”
“I made thousands selling a book made by AI.”
“THEFT! EVIL! SOMEONE NEEDS TO SDO SOMETHING WITH AI!!!!”

There was another comment further up that said something to the effect of “He’s only doing it for the money!” But like, look around you, people. Literally the entirety of modern society is based on people doing things for money. And I’m the bad guy twirling a mustache while a beautiful damsel labelled “writing” is tied to the train tracks? I think not. Or, at least, don’t stop with me. Carry that criticism all the way forward.

And this has always been a point of mine:

I don’t care.

Are the books good? Because there are a lot of books out there, written by humans, that are garbage.

Plenty of humans chase the algorithm. Plenty of humans churn out garbage based on trends, and chasing pennies. Judge the work by the work.

Is this true about Tron?

If people enjoy the books who cares, so ridiculous. When Tron wasn’t eligible for an academy award because computer effects were “cheating”(look it up). Every time something new comes along this happens, it happened with PC paint programs when they started to get big, autotune was just “fake” music. If its entertaining and sells i don’t care where it comes from.

According to Wikipedia, it’s true! Who knew! (Apparently this guy ^^)

Tron received nominations for Best Costume Design^[2] and Best Sound^[3] at the 55th Academy Awards. It was however disqualified from the Best Visual Effects category because at the time the Academy felt that using computer animation was “cheating”.^[4][5]

There are also definitely some AI-generated comments on this thread, but they’re boring, so I won’t reproduce them.

This idea that all of this is somehow “political repression” is interesting when we think about how integrated technology is now into politics, culture, etc.

An artist should be free to choose their own tools and media.

Anything else is political repression of art.

Don’t like it? Don’t buy it.

I think there may actually be a more even balance of semi-supporters on this thread than I have seen previously, so perhaps the tide is starting to turn…

Why does it matter how people create work that others enjoy? Next people will complain that he didn’t write the books on a manual typewriter or use parchment and quill.

That’s all for now. Almost caught up on my backlog here, but a few of the backlog items are a lot bigger and deserve one or multiple posts each, which I’ll continue to plug away on over the coming days.

I’ve gotten interested in this recurring criticism on social media of my work with gen AI, that by using these tools I’m “not a real artist.” So I’ve been looking at other new and interesting ways I could provoke that same criticism, all while thoroughly expressing my not-real-artistness.

I landed on a simple first experiment after looking for videos of drawing machines and robots on YouTube. We made this one at home from a spare battery pack, a motor I pulled from a broken toy bubble gun, markers, etc. When the alligator clip gets connected, current flows, and the bot hops and bobbles around on account of the motor’s attached weight being off-kilter. Video here on Imgur, I can’t seem to embed it into this WordPress install.

The doodle bot outputs can look something like this:

This one contains traces from several successive runs of the machine on the same paper, and here’s another on some scrap that was previously marked up with black paint, as depicted in the linked video:

I could easily imagine more developed experiments in this direction which, combined with maybe other painting techniques, and using some acrylic paint pens instead of felt-tip markers might actually result in some pretty compelling finished works.

But are they copyrightable? This is another common concern people seem to have in reaction to my work with AI. I tend to follow conceptually the UK copyright framing, as described here:

Computer-generated works

The UK is one of only a handful of countries to protect works generated by a computer where there is no human creator. The “author” of a “computer-generated work” (CGW) is defined as “the person by whom the arrangements necessary for the creation of the work are undertaken”. Protection lasts for 50 years from the date the work is made.

For my money, that’s the most rational and reasonable policy regarding copyright for computer-generated works, and it’s quite divergent from how the US Copyright Office is currently handling these matters.

In poking around for historic antecedents of drawing machines, I came across some cool work by the Swiss artist Jean Tinguely, whose work in “méta-matics” is depicted here in this short 1959 British Pathé video:

Tinguely’s work is too deep to really go fully into in this blog post, but here’s another video to give further context:

Worthy of more exploration another time. It’s cool that he was also concerned with the poetry of the machines and their actions and relations themselves, in addition to this ability he sometimes experimented with, of having them output drawings as records of their erratic motions.

Of course there is Harold Cohen, whose work I briefly went into recently here. The interesting thing about his work is that it was an attempt at a rules-based AI, not statistics-based like today’s gen AI. So he hard-coded all these physical relationships and proportions of the human body, so that his machines and software could autonomously generate consistent and aesthetically pleasing visual artifacts.

Sougwen Chung is also integrating more sophisticated drawing robots into her work, as I went into a little bit here.

Another prominent and more recent example I found in this space are the drawing machines of Harvey Moon, which are captured in this video from 12 years ago:

That video depicts a wall plotter, which also goes by the name of polargraph, and many other names for more or less the same thing, a motorized mechanism connected to software which can shuttle around a marking device against a drawing surface on a wall or panel.

There are tons of tutorials about how to build one of these things, and control it usually with Arduino to trace out shapes defined in graphics software. I thought this one by Liz Melchor was one of the better ones. If you pop over to this animated gif on their site, you can see one of those things in actions, complete with counter-weights on each side.

It occurred to me somewhere along the way though, couldn’t I build something like this but without any motors or software? Essentially a “human-powered” extremely low-tech drawing “robot.” So I built one in my basement with some rope, pulleys, part of a plastic bottle with some coins in it as a gondola, and voila (click through for a short video of it in action, not able to embed here):

Here’s a schematic I did that is a reconfigured version of the polargraph schematic originally found here.

My version replaces the motors in top right corners with simple pulleys, and the counter-weights with another set of pulleys below to be able to operate the mechanism from a few feet back.

This is obviously a quick and crude prototype, but it tells me that with some tinkering and refinement, there is some interesting territory here to be carved out. I couldn’t really find anyone else who was doing anything like a manual wall-plotter.

I’ve realized about myself after all these years that, even if I can occasionally pop in and get some small bits of code working, I’m really not a “coder” and I don’t enjoy getting bogged down in trouble-shooting Arduino, etc. So the idea of simply side-stepping all the high-tech elements and subverting them with low-budget low-tech alternatives powered by human “elbow-grease” sounds really appealing to me. I worked for years as a stage technician, so this kind of simple rigging to perform an action in the moment makes a lot of sense to me.

I did have someone on Reddit point out that a large scale manual wall plotter could also essentially be a large wall-size Etch-a-Sketch which would be really amazing to have. This video shows what those devices look like inside, how they work, and offer a crude attempt at duplicating the mechanism:

The Etch-A-Sketch Wikipedia page has an excellent diagram of the pulleys and rigging inside the machine and a verbal explanation, which I’ll include below for fun to show how it shares many elements we might see in a CNC plotter:

I’m planning to build one of these to the dimensions of my experimental wall, but in the meantime, it’s interesting to see in my first tests of the basic prototype above that certain kinds of movements are easy to pull off, and certain are more difficult, and it has everything to do with the points of control and the methods of their functioning.

In my survey of this landscape, I also found an excellent site dedicated to various types of historical drawing machines. One of the more common of these that are easy to find on YouTube is the harmonograph, consisting of anywhere usually from two to four pendulums. This one has three:

I’m planning to build one of these under a piece of furniture I made recently to do a test. It will be a variation of this design:

Am I a “real artist” yet or not? I can’t even tell anymore. I think the question *might* just be meaningless…

There’s another direction to explore in the purely mechanical drawing methods, that of using irregularly shaped cams which correspond to repeatable drawn patterns, as can be somewhat seen in this two videos from Aaron Kramer:

That looks like a really fun and interesting and time-consuming thing to get into… Putting that on the list. Another video about something similar with 3d printed cams by Takuto Takahashi:

This Carnegie Mellon University Design video has some other interesting random drawing machines of many different types:

And an interesting paper entitled A Brief Overview on the Evolution of Drawing Machines, which includes a nice set of base criteria for what might constitute a drawing machine in the first place:

1. A drawing machine can be an autonomous or semi-autonomous machine. This can be a set of human-powered gears or mechanical linkages that automatically generates an image through a machine-held stylus.
2. A drawing machine must control—or help a user control—a stylus, a pointy object that leaves a mark or line on a surface when applied pressure: pen, quill, pencil, airbrush, or more recently capacitive tips for touchscreens.
3. When used to draw from life, a drawing machine inserts itself into the stylus-hand-eye circuit. As the artist holds a stylus in her hand, whose movements are coordinated by eye, the drawing machine can guide the eye, or control the stylus, or augment the hand.

Maybe there are other criteria I’d like to explore in this “latent space” but those seem like good jumping off points for further exploration.

Speaking of human-powered machines, I felt it also might be relevant to get briefly into automatism or automatic drawing, especially as pioneered by the Surrealists. Perhaps it could be said that they were trying to turn their bodies into robots which could express the hidden non-verbal commands of the subconscious mind through them as mechanism. Lots more to think about here!

I haven’t had a chance to update the English ebook version yet, but the French print edition of The Quatria Conspiracy has all new art in the style of classic era black and white pulp sci-fi/comic illustrations. Here’s one below, and at the link you can see a bunch more.

These are made in Dalle, but I believe the print edition contains images from other sources as well.