Talk me off the ledge: Will AI destroy my industry?

That’s a reassuring POV, but if it were just AI copy vs. my copy, I wouldn’t be so worried. The article I cited described someone basically replacing an entire agency team with cobbled-together AI tools. I can just see a client thinking, Hmmm, I know the AI won’t deliver a campaign as good as Akaj, Inc. will, but I can tweak and polish whatever it comes up with and save myself $30K and a month’s development.

If the resulting work underperforms, they may not realize their mistake until we’ve been fired and I’ve been laid off.

On the other hand, you’ll have a hundred times the people now who can generate tons of fanciful images from their head who previously lacked the time/skill/software/tools/talent to do it the previous way. If your goal is to make D&D character art or even just visually interesting stuff to hang on the walls of your living room, then it seems like more of a positive than a negative to give that ability to many times more people.

Not great for the current crop of artists financially but probably a net gain on human creative output.

The software development industry went through exactly this problem for the last ~20-25 years.

But the opposition then was not AI. It was outsourcing companies in low-wage countries, primarily India. Lots and lots of truly awful software code by truly awful overseas workers was done while US execs salivated at the cost savings and laid off large swathes of their own techies.

The chickens began coming home to roost almost immediately, but plenty of bosses got their bonuses and changed jobs before the steaming piles of crap fell apart completely. Many of the cost savings proved illusory once they had to pay two or three times to get the job done close enough to correctly to actually deliver on the promised business results.

Fast forward 20 years and now a) the quality of the best offshore work is at least good enough, b) there is still a vast amount of total shite produced for customers who do not care, c) managers are a lot more wary but nowhere near wary enough, d) many US jobs that used to be “Do the tech” are now “Monitor the foreigners doing the tech”, and e) there are far fewer "monitoring " jobs than there were “doing” jobs on any given project, but the overall industry has expanded enough to absorb most of the people with an aptitude for both roles.

You can sort out how this applies to your industry and see where you should move your tent towards so it’ll be standing on firmer ground rather than quicksand.

This is true.

Then again, it’s equally true of astrology, horoscopes and clairvoyancy. And vast swathes of political spin and PR. Indeed most of the literature which crosses my RL in-tray every day.

The post by @Shalmanese is exceptional. I dips me lid to you. :mortar_board:
Do they own the content? Or have they simply “retweeted” ChatGPT? Are they responsible for correcting any factual errors? Does ChatGPT deliver facts, or discoveries or simply opinions? If the content has adverse and dire consequences, who issues the clarification, correction or apology?

Because if it is simply delivering bromides to allay my concerns, it did a rather good job. Except that this concerns me.

Is it, though? Plenty of commercial art today certainly uses computer-aided tools to produce art much faster and with greater detail than would be possible by laborious hand-drawing processes, but this is in a real sense replacing the ‘artist’ with a ‘art director’. It is certainly possible to tell a gerative ‘AI’ to give you an Escher-like creature, but then, what becomes of the impetus in formenting the next Escher? Setting aside the safety concerns of how this capability can be misused (which I think are quite tangible), the effect of suppressing creativity and innovation in pursuit of output volume.

McDonalds may have served “billions and billions” and they’ve certainly put a lot of effort into both marketing to consumers and manipulation of consumer neurological impulses to maximize their profits, but they have not advanced the culinary art of the hamburger and have demonstrably harmed society in a multitude of ways. I view these systems, used injudiciously not to just enhance or amplify and improve human creations but to replace human artists almost entirely in many commercial applications as the creative equivalent of fast food any with many comparable negative impacts.

Stranger

A long time ago, half a century ago, I was at school in a town where the local paper had three editions. On Monday the back page was filled with stories of the past weekend’s footy games. On Friday the back page was filled with stories and predictions of the upcoming weekends footy games. But on Wednesday the back page was filled with individual match descriptions of the previous weekends games of the town’s second sport, (field) hockey. 10 column centimetres, maybe more on each of the six 1st grade mens games, a bit less on the 2nd, 3rd and juniors fixtures. The womens competition was roughly the same scale. Written by one guy. But the games were played at two venues, in three time slots. The editor played in one, usually umpired another. Would only have seen even obliquely say 10 of the 48 games played. The other details came from the scorecards, names, positions, goals, best players and maybe a chat to an umpire or captain of the 1s games. But if you didn’t know, you couldn’t guess which ones Shep played and which ones he remotely reported.

I have not heard it expressed in these terms, but damn that’s apt! Well done!

I think so. When I’m in a Discord channel with friends and we’re riffing off one another and prompting all sorts of stuff, refining it and using one another’s ideas I definitely feel like there’s a creative process going on.

This is my fear. The people who hire us will reap the promotions from delivering campaigns faster and more cheaply than ever before, and by the time their companies recognize the shortcomings of that work my agency and hundreds like it will be defunct.

My other fear is that the AI-generated campaigns won’t be inferior.

If I were you or your company, I’d jump in with both feet. Imagine what you could do with this tool as it evolves. If it’s not inferior to what you do now (unlikely) imagine what you could do with your current product as a starting point!

The early days will favor the bold more entrepreneurial workers. Soon @Akaj working alone may do the work of two floors-worth of her now bankrupt former employer’s now unemployed former employees.

If @Akaj hopes to remain a safe corporate droid, punching out content like a metronome for 40 hours times however many more years, that way lies unemployment or prompt (and perhaps too soon) retirement.

Surf the wave early; that’s where the best hope of success lies. Eventually some dust will settle out and organized agencies will once again out-compete solos; scale does give power. But by then somebody who is known in the brave new world can be a bigger player in that new agency world.

On the highway of progress, the slow are roadkill, and the fast get to join the club.

At my job- a video game development studio- we already use AI to generate concept art. We still employ concept artists, but now their jobs are primarily in generating these pieces and then doing paintovers on them to make sure we get exactly what we need. It’s a massive timesaver- and the obvious corollary is that we won’t need as many concept artists in the future.

However, I have no illusions that my job- special effects- will not someday be fully AI. Luckily, I’m only about ten years from retirement, so it’ll probably be someone else’s problem when that day comes.

I seriously thought that we’d see the more manual labor jobs replaced by automation well before jobs like mine would be made obsolete, but I don’t think that’s the case anymore.

This may be an overly fine distinction, but when you are working with a group of people, you are engaged in a communal or at least interactive creative process that draws off of each contributor’s personal experience. When you are using a generative AI to make images or gin up some text, you are just feeding it some parameters and getting a result. There may be the impression of collaboration but I think there is a real if subtle loss in exercising true skills of creativity, and I am very concerned what becomes of the generation of people who have had this at their fingertips (or more likely voice, as the keyboard is likely the next interface to become arcane and unneeded).

Now, maybe I’m just a fogey and don’t understand all of these new-fangled TikerTokers with their dance moves and face filters much less those who will come after them, but I think there is a qualitative shift in tacitly handing over too much creative faculty over to a ‘thing’ that does this with virtually no effort on the part of the user. I’m far less worried about killbots and computer overlords (although I am worried about drones and cyberautonomy) than I am of humanity essentially handing over its collective will to an AI overseer that we don’t really even understand or can predict its path. And while it is true that the future has always been full of unforeseen technological changes that shape society in unpredictable way, this is a force that can take an active role in doing so in ways that are not necessarily in the best interests of humanity overall.

We aim to please, and missing that, to at least amuse.

Stranger

Sure. And you’re creatively coming up with said parameters. Coming up with ideas, Mixing terms, descriptions, various artists and styles, etc to try to realize those ideas, adjusting that, editing results via inpainting (depending on the client), etc. I’m not going to debate if it’s “real art” because I don’t care but it’s definitely a creative process and feels weird to suggest otherwise by gatekeeping “true creativity” behind paintbrushes and digital art pads.

I’m no spring chicken but I assume this is a big part of it. Old people, new tech, fists shaking at clouds about the olden days and how this will make all the kids soft and lazy.

Above and beyond that, I’m not remotely worried about it. People have been involved in creative expression since we were scratching sticks in the dirt and it’s not going to stop now just because they found a new way to make it happen.

Just to be clear, I’m not saying that generative AI works are not art; they are clearly products of craft (although not craft actually developed by the AI; they are amalgamations of ideas and creative skills of others, stitched together in a trained synthesis), and some of them are certainly ‘unique’, albeit in ways that are more interesting than appealing without a lot of help from human direction. And I’m not at all opposed to the use of computer-aided systems to enhance what a human user can do in terms of both the speed and sophistication of output; for sure, these will increase ‘productivity’, and in some ways may create new jobs or even entire new fields of endeavor. But these systems are qualitatively different from digital art tools because they are now both co-oping the collective craft of many others without conveying that skill to the user.

As a comparison to my own field in (occasionally) doing structural analysis, going from hand calculations and empirical methods to high fidelity finite element analysis is an unambiguous improvement that allows me to gain a far better understanding of details stresses and strains, and the ability to iterate through different configurations and parameters to optimize a structure in various ways is a remarkable leap forward that in no way removes my agency or demands less knowledge and intuition from me. But once you go into ‘automated’ FEA tools that hide the mesh and start doing a bunch of ‘optimization’ without consideration for the practical knowledge such as how real worlds structures are built or why a certain type of modeling idealization will produce false results, I’m now in an area where I no longer fully control or understand the tool and may end up with results that are suspect even if the system can do the analysis much faster than I could in building the model geometry and mesh myself. And knowing that, I have a feel for just how far I can trust such a system to give me a useful result. But someone who has not come up through having to think through structural problems and do calculations by hand, nor has built up a mesh or thought very deeply about the impact of idealizations may just blindly trust the system to build a good model and lack intuition about when the results are not as expected.

With a physical structure, the potential for a bad analysis is obvious; it may fail, possibly catastrophically. With art or literature, the detrimental effects may be more subtle (but may also be dramatic as AI learns how to directly manipulate human nature and subvert even the patina of volition that you think you employ in decision making) but that makes them even more insidious as people lose the ability to create or even think for themselves writ large. I don’t say this in the abstract; we have empirical and observable evidence of what social media has done to critical thinking and moderation of emotional outrage, and chatbots and other interactive and generative AI have the potential to take that to an entirely new level. I sure hope I’m wrong and just don’t get ‘it’, but I think I do understand the risks with great clarity, and I’m very concerned about what bad actors, or just the general desire to do things the easy way, will do to human society and intellectual potential.

Stranger

Depends on what you consider to be “the skill”. Is it handling a paintbrush? Or is it creating ideas and concepts and conveying them? Can you be creative with a sticker book if you didn’t design the stickers and are just placing someone else’s work on a page? I think what’s going on with AI prompting exceeds that but then I also think someone can absolutely be creative with a sticker book so it’s a moot point to me.

I guess I’ll leave it there because I don’t see me convincing you and, ultimately, I guess it doesn’t matter to me if you end the day convinced that AI Art will sap the human creative spirit dry. Completely disagree with you though.

First of all, I’m a dude.

More important, there’s a helluva lot of middle between being a solo act and a “safe corporate droid.” I’ve done the freelance thing and vastly prefer the teamwork, creative interplay and, yes, affordable health insurance of agency life. While I could probably invest the time and effort to master the AI-wrangling I’d need to forge the kind of entrepreneurial success you describe, I don’t want to ditch my colleagues and run my own business at this point in my life. And I sort of resent the suggestion that feeling this way makes me a soulless “metronome.”

I’m not throwing up my hands in surrender; in fact, I’m spearheading my team’s efforts to leverage these new technologies into improved client offerings, and I think we have a chance to make it work. But it’s also just as likely that client cost-cutting mentalities will cut us off before we have a chance.

…if it’s your company that is “jumping in with both feet” then the likely result would be you getting fired and being replaced with someone that will do the job cheaper.

I’d take an experienced copywriter over an AI (as it is right now) any day of the week. They don’t need to use AI as a starting point. What the client needs and wants is the starting point. It’s very easy to write generic copy that can apply to basically any business that works in a specific industry. But it takes skill to write something that genuinely reflects and represents the brand you are writing about.

I’m now an independent web designer. I use AI sparingly. It’s good for automatically generating alt-text: AI is good enough now it can look at an image and provide a not-that-bad description of the image that I can tweak. This is a HUGE time-saver. Especially when dealing with hundreds of images.

Its also not bad as a starting point for meta-titles and meta-descriptions. But even something as simple as a meta-title requires a degree of tweaking before it goes live.

And I also use it when my brain stops working. Sometimes I need to write something and my brain just doesn’t want to help. So I’ll play with prompts until I get maybe a sentence or two that kick-starts my brain, and then I’ll take it from there.

But as a small business owner this isn’t something I’m going to jump into with both feet for a very simple reason: I offer very personalised services and AI is the antithesis of that.

In general, where do I see the industry going?

It will go in the very same direction as everything else that the tech-bros have gotten their hands on, from everything like Uber to DoorDash to Airbnb. It will drive wages down. People with skills will leave the industry. The products and services won’t get any cheaper. The already rich will get quite a bit richer. And things will get just a bit more generic, a bit more soulless, a bit more blah.

Niche, independent small businesses will likely be relatively unaffected. I only need between 12 and 15 clients a year to be profitable, and I’ll pretty much always be able to find that, somehow, some way. The biggest change I think is that a lot of businesses take things in-house, which isn’t necessarily a bad thing, just an evolution. The great creative agencies and use AI as a tool to enhance what they already do. The bad creative agencies will use AI to be more exploitative than they already are.

Things that I’m doing right now is that I’m creating an AI disclosure document so that people know to what extent AI is used in the creation of anything I do. But AI isn’t something that I see having a significant impact personally on my own business.

I’m curious if you read the Wharton article in the OP. If you (as a creator of any sort) could start with that and many different iterations from different seeds why would that not be a huge advantage? It will definitely put people out of work, but for those “agile” (not Musk) to use it it seems huge.

It isn’t so much that I think that “AI Art will sap the human creative spirit dry”—I don’t think that is possible because there will always be people inclinded toward creativity—but that generative AI will become so available and inexpensive that there will be little motivation to pay for original art (from a commercial standpoint…there will always be people ready to ‘invest’ in highly regarded ‘art’ in various forms) and little motiviation for people to pursue art at a vocation. I fear even more for writing, and especially fiction because writing fiction is already a process of synthesizing experience, knowledge, and inspiration from the work of others into mostly genre-defined forms. Art and writing are of the few things that both distinguishes us from other animals and that bring us together in a shared connection that goes beyond personal bonds, and to have that subverted by an generative machine ‘intelligence’ that can do it almost as well for most purposes diminishes not just the uniqueness of human culture but the motivation to advance and improve.

But I am not the ulatime arbiter of what is right, nor do I hold any exclusive knowledge over what may come. For all I know, generative AI may spur a competition with human artists and writers to produce works that are more unique and innovative just to distinguish their work (although that comes with its own problems). I think we can say with confidence that generative AI will certainly be adopted by companies and contractors to aid them competitively, and whether you regard that as good or bad depends upon your outlook, although I doubt novelty and creatively will be on the forefront of consideration.

This would be the sensible way to use these tools; as an aid and starting point. Unfortuantely, I doubt that many people will stop there because it is just too easy to rely upon these tools, and as they get better the distinction between generated text and that produced by a person will become narrower, to the point that unless you are producing something really unique, it is probably good enough. It’s like making a modern pop music star; it doesn’t matter if they have a weak singing voice or severe emotional problems, so long as you Auto-Tune the tracks and keep them under control for long enough to cut an album and then send them to therapy/rehab/remote private resort until the next hot thing comes along.

Stranger