The Resistance

How do we best protect our works from being abused by AI?

How can small time writers fight in the resistance? What can we do? Keep all our work offline on a disconnected computer until it is ready to publish?

Not all of us can fund legal battles, will there be a fund for human writers to help them?

There might be resources to help us, such as AI identification programs, but they may not always be reliable.

Some things we can do ourselves:
-Rely on human artists for our cover and illustration work.
-Do not upload your stories digitally, make sure they only exist in printed form. (Might be impossible and a shot in your own foot.)
-Avoid using/paying for commercial products that have allied with AI. (Might become harder and harder.)

Or do we succumb to the lures of easy creation?

Any ideas? Resources?

 

There are a couple of interesting ideas.

It is possible to poison the AI training algorithms and screw them up. There is a tool online that adds a layer of noise to your digital art. Any AI that then try to read the image fail to because of the noise and they effectively learn nothing from it.

The technology to do this is called Nightshade. It will be part of the Glaze tool but it doesn't look like it's available to the public yet. The potential looks promising though.

And it is only for images. I'm really not sure something like this is possible for written words. But it is an interesting proof of concept. You can read more about Glaze here.

I truly hope that more tools like this will come out. OpenAI has no moral compass so defence is entirely up to us.

 

There are class action lawsuits one could join. There are also places like the Authors Guild that will help writers with this issue.

 

I was told not to save anything to 'a cloud'.

 

Yes! Big tech companies like Microsoft and Google cannot be trusted. The files you store in places like OneDrive and Google Drive is something they have full access to. Of course, they will tell you that they'd never 'access' your data, but you have absolutely no idea what they do behind their closed doors. Their interest is making it big while generating big cash in the process. Your privacy is the least of their concern.

There is nothing stopping Microsoft from scraping Word files inside OneDrive accounts and then training their models with them. Nothing.

As a matter of fact, the Mozilla Foundation is currently in the process of trying to find out just how much data they're harvesting from their eco-system (Microsoft Word, Windows, OneDrive, Teams, etc.) to train their AI models. They have nothing in their privacy policy to state whether they do this or not. Considering their track record of invasive tactics and terrible privacy, it wouldn't be far-fetched to say that they're.

So... it is very possible that the mere act of using Microsoft Word generates data for their AI. Maybe not your entire novel but parts of it. Or maybe it is your entire novel. That's the thing. We have no idea.

This is exactly the reason I distanced myself from Big Tech as much as possible. I use Linux and LibreOffice on my computer and GrapheneOS on my phone. All of these can be audited and reviewed by anyone, including your and me. There is no hiding anything.

Here is the Mozilla Foundation's petition about this issue. The more people sign it, the more pressure we put to Microsoft.

 

So informing the consumer might be the best option other than protecting yourself? How can the writing and arts community best inform the public about these issues? Mention them over and over all over the place? Might be a good start.

I hope we get "Human Made" quality stamps that I briefly mentioned in another thread.

I wonder how this whole AI situation can be balanced out. We may need it for science and industry so we don't risk lagging behind other potential civilisations. But it should be banned from partaking in creative fields in my opinion, for now.

 

I think the novelty may initially counteract playing to consumers' desire for 'genuine' product. "You mean a robot fried this rice? Cool."

And on top of that, I just don't think ethics drive a lot of spending decisions. It's a mix of convenience and cost. We're treated like ungulates because we act like them. Before you say: "Well I don't purchase Nestle products because of the such and such..." You're honourable/special/conscientious, but you're the minority.

AI will find some sort of place, but I still don't think it will hurt authors as much as it will, say, graphical/musical artists. Huh, another group of people that make barely any money these days.

 

I think Bud Light would disagree with you on this point. It is not about virtue signaling. There are a number of companies whose products I avoid, and look for something else to fill the same need. It is about the moral/ethical disagreement, not about talking about it. Virtue signaling is calling attention to yourself, saying look how good i am for taking this position.

In the end, "contains AI content" will end up being like "made in China" on a product with similar quality issues.

 

I meant that more as in someone will reply that they do in fact take ethics of consumer products into account with their purchases as a common counterpoint (well, I use my wallet conscientiously), but they are inherently not the majority whether they talk about their virtue or not. For every instance where Bud Light's pop culture PR has hurt its profits, there's an Iphone, a chocolate bar, a bottle of water, an article of clothing, an accessory, an electrical appliance, a TikTok, a search engine etc... that was still consumed or used regardless of its company's ethics. The evidence for this is how successful these corporations still are.

 

And it's not even just politics, these days companies have a tendency to backstab their own best customers (though it's usually because of politics, or ESG score, which is politics).

The local classic rock station recently hired a youg DJ, I think they're realizing they need to appeal to a younger audience, because us old coots will be dying out. Almost all of their current DJs were people who had been with the station for decades, as old as or older than me. Well, this new guy is an ass clown, acts like a shock jock morning host, but they put him on the weekday afternoon shift, and he likes to insult the older listeners like me. The station is called K-SHE. Interesting story—it was originally formed in the early 60's specifically to be a feminist radio station, but I guess it failed, and a rock station bought it and kept the call letters and the name. But they spell everything with a K at the beginning, like on Sundays they have a program called Kshe Klassics, and they call the more bizarre people who call in Kshe Krazies. Which is also very new, and goes hand in hand with what I'm talking about. Anyway, this new DJ calls me and people like me Kshe Karens, because we complain about him and his lack of respect and professionalism. His idea of humor is to insult people, and he likes to insult the listeners as he's talking to them on the phone, on the air. He's pissed me off so much by now I don't even listen while he's on anymore, with all his nasty insults about "All these old guys who are all like "FREEBIRD!"

Yeah, like we're all dumb hick rednecks or something. What really bugs me is that the radio station lets him do it. What ever happened to not biting the hand that feeds you, or to professionalism? You don't insult your biggest customer base and expect to survive as a company.

It's as bad as the marketing director at Anheuser Bush saying their customer base was too 'fratty' and male or whatever it was. Ok, fine, all us fratty middle aged guys feel the same way about you. And now who's gonna buy your beer? Oops...

 

That's actually wild about them getting you to pretend purchase Bud.

Well, sure, we can talk about A.B, but that IMO stands out as the demonstrating exception as far as boycotts go. My comment was in regards to the less sensational things that companies get away with every day. A cross-promotion with a trans influencer =/= using tertiary slave labour or other international exploits, but it ends up being more significant in the public eye. AI will never be as controversial as Dylan Mulvaney, not even close, even if it's leagues more important, so don't count on the consumer spending with respect to writer employment.

On A.B. in particular: I think the backlash was taking their marketing robots a bit too seriously. Fun fact, trans people are apparently about twice as likely as the average person to have a drinking problem. Who's the audience? Makes me wonder how innocuous the move really was. Either way, just do a funny commercial with a cute dog in it. Worked for ages, will not stop working.

 

Random thought.

If you write badly enough, the machines won't be able to copy your work because they will not be able to understand the grammar. Yet people might be able to discern your texts because they dont think like a machine.

Idiot-genius way of AI-proofing your work? Take a step back in your grammar development. One could write a novel in fragmented sentences and such. Heh, maybe not...

 

Ha ha—write like James Joyce, or dude who wrote A Clockwork Orange. Nothing but intense street slang.