Can journalism go on strike?
No but it would be nice.
I’ve been trying to write this post for a few weeks now, and there’s no good reason why it’s taken me so long. Sure, a lot of my procrastination had to do with some well-earned burnout, but there’s something unwieldy, intimidating, and seemingly useless about arguing for labor. As the title of this post suggests, I’ve been wanting to write about the possibilities of journalism going on strike. But in thinking about that for more than 10 seconds, the idea starts to fall apart. There is no focal point for a national movement, no organization that can take the reins, and no large way for freelancers — who are often not in any union because they don’t exist — to be represented. Plus, the media is international; if American workers go on strike, the companies can just go abroad and ignore them. And as I’ve seen with countless publications, you can get away with paying people in certain countries a lot less money than you can in the U.S.
I wanted to continue with the topic though. There’s something in the water that’s been fueling a lot of workers this Strike Summer, and it just feels like a good time for journalism to have their turn.
Labor movements have been on the rise over the past couple of years — since the pandemic really — but this summer we’ve seen a number of huge strikes and strike authorizations across multiple industries. The most prolific this season is the WGA strike, which was soon joined by SAG-AFTRA and shut down Hollywood. After many damning reports on the state of VFX work on blockbusters like the MCU, VFX crews at Marvel have petitioned the National Labor Relations Board (NLRB) to hold a vote to unionize. The UPS union was also on the verge of a strike before a last-minute deal with the Teamsters gave the workers what they asked for. The list goes on.
What struck me the most about these strikes (besides how cool they are generally) is what many of these groups were fighting for:
Increased compensation for all workers, which includes contractors and part-time employees, along with better benefits, protections, and updated compensation for the streaming era.
Provisions against getting their jobs replaced by generative AI tools.
The latter obviously stuck out to me as somebody who’s been following the extremely quick adoption of large language models (LLM) (colloquially known as AI) in media newsrooms. I still stand by the idea that the technology isn’t the issue as much as the people who want to utilize it, but the tech has also continued to prove it’s just not ready for much beyond basic assistance for spellchecking or creating images for your personal role-playing campaign. It’s impossible to use in a professional setting without immense risk, whether you’re using it to create actual art to be sold, share information without accidentally revealing a lot of internal secrets, or create articles for publication. You will either end up with incorrect information or a plagiarism claim, and none of that sounds appealing.
And yet, despite so many instances of major companies getting caught using AI in the most embarrassing ways possible, they still continue to push forward initiatives that want to use the technology and replace the work of actual people, regardless of what the PR says. CNET continues to ignore how its AI tool will plagiarize even after many months of public fuckups. But most egregious recently came from G/O Media, which had transparently noted it was going to start using AI across its websites and then just started doing it without alerting editors.
I first noticed an article on io9 back in early July about the Star Wars movies in chronological order, which was already strange to see considering the caliber of the writers and the sheer amount of Star Wars knowledge available. But clicking in you realize that not only is it lacking any structure or human touch — it just hops into the list without introduction or any creativity — but it was also just incorrect. For one, it didn’t include every Star Wars property. Then it got the order wrong! It was the one thing it had to do correctly, and it failed at that.
It then became apparent that this wasn’t an isolated incident. Articles from “bots” had also appeared on Deadspin and The AV Club. At best, they were filled with unoriginal drivel. At worst, they were flat-out inaccurate.
The io9 article was edited after it started to go viral, but accuracy isn’t the point. The clicks are the point. Nobody’s going to read these articles, and that suits executives just fine. The hope with articles like this is that they will get on the first page of Google and get a couple of clicks before people realize it was written by a bot. Even better, these companies want to populate the bullet points in the featured snippet box at the top of search results. Since Google introduced this feature in 2014, publications have not only been fighting for the front page but also that top box. And what Google loves for that box are bullet points and numbered lists. Most sites just changed the formatting on the articles themselves to placate Google. G/O decided to just throw out the “article” part.
And in the case of the writers’ and actors’ strikes, the AI is literally replacing people. Ahead of the SAG-AFTRA strike announcement, the Alliance of Motion Picture and Television Producers (AMPTP) said in a statement that it had offered a “groundbreaking AI proposal” that would protect “performers’ digital likenesses.” However, in reality, they proposed something so obviously evil and against what the unions were working for.
According to Duncan Crabtree-Ireland, SAG-AFTRA’s chief negotiator:
”They proposed that our background performers should be able to be scanned, get one day’s pay, and their companies should own that scan, their image, their likeness and should be able to use it for the rest of eternity on any project they want, with no consent and no compensation. So if you think that’s a groundbreaking proposal, I suggest you think again.”
The AMPTP denied the claims, but the writing is on the wall. Actors’ physical likenesses and voices are already being digitally replicated. Star Wars has been particularly egregious with its use of dead actors and uncanny de-aging; the recent Ghostbusters movie resurrected Harold Ramis; Warner Bros. Discovery recently decided to resurrect Christopher Reeve in The Flash for an unnecessary seconds-long cameo. Some of these were done in service of a story’s timeline, like with the appearance of Peter Cushing in Rogue One, but the ethical implications of recreating a dead actor just for some box office return or streaming numbers shouldn’t be up for debate, even if it sort of helps with streamlining canon.
And if the studios can recreate dead actors for blockbuster shlock, why not do it with background actors or ones they don’t have to pay multiple days for?
Despite the fact multiple Hollywood executives said they were waiting for the strikers to run out of money and start losing their homes (yes that is real), they seem to now want to reopen talks. Most recently, the WGA and the AMPTP met for the first time since the strike began, (although the WGA reported that the AMPTP leaked the contents of the meeting to the press, which went against their media blackout agreement). At this point, we don’t know if the unions will get everything they want, but maybe blocking productions, not promoting new releases, and generally not working is making the AMPTP actually pay attention. And as we can see with the UPS strike that never was, even just the threat of a strike can do wonders for putting pressure on negotiations and just general worker morale.
I know this isn’t the most nuanced take. I know there are countless reasons why people would be against striking that aren’t necessarily anti-labor. It doesn’t help that the National Labor Relations Act, a huge piece of legislation that gives people the right to unionize, doesn’t do nearly enough to hold companies accountable for anti-union activities. Why unionize when it’s just going to make everything worse or if it’s going to amount to nothing? (Looking at you, Amazon and Starbucks.) Still, I don’t think the conversation should be complicated. You put pressure on those with the power and the money, and you can force their hand.
Can we hold generative AI companies and others who would like to pay for it accountable? That would be nice.
Take, for example, the fights for railroad workers last year. Sure, the Biden administration did not handle the pending strike well, but it did eventually lead to talks that got the unions what they wanted. According to the Association of American Railroads, which represents freight rail workers, since those national talks concluded, “most unionized employees… are now covered by agreements providing individual paid sick leave days,” which was the unions’ biggest negotiation point. The national coverage ended when Biden blocked the strike, but the actual talks did not.
So maybe now is the time for journalism to do… something. I’m not sure what. But things are only going to get worse, especially when it comes to generative AI. Bing and Google have already been pushing their AI chatbots — Chat and Bard, respectively — into their search engines. Granted, search engines these days are useless and desperately need overhauls anyway, but LLMs aren’t the way to do it. They can be neat little tools if you want to kickstart research or want to quickly pull up a few recipes, but they’re too immature and literal to do any professional work where accuracy and context are critical.
If I go to Bing Chat and ask it to tell me who I am, it can give me info it pulled from a few different websites, but it doesn’t account for the passage of time. Previously, it had told me I still ran Postmortem Mag, a project I did for around a year back in 2014 or so. I also asked it a couple questions this week and it still says that you can email me at my last job, where I was laid off almost a year ago.
Even more terrifying is that according to The New York Times (archive link), Google is testing AI tools that can actually write news articles. Considering AI’s great track record in crafting accurate articles that people will want to read (see above), I see this being an excellent way to collapse the whole industry.
Think about it. Generative AI models are trained on content on the internet — photos, videos, art, words, anything. If Google and other outlets output art and information using AI, and that content gets pushed up in Google results, all we’ll have is AI-generated crap. Then, the models will use that content to train itself, creating an endless loop of crap filing into other crap to create an internet that’s only full of crap. If journalism’s primary goal is to keep readers informed of the world, that sounds counterproductive.
Weirdly, one of the biggest institutions fighting against the proliferation of AI has been Getty Images. Most journalists know them as the place to go if you need just about any image for your article and don’t mind paying $500 for it. But, as you can imagine, AI is going to cut into its business. Regardless of motivations, it filed two lawsuits against Stability AI, the company behind the Stable Diffusion image generator engine, earlier this year, claiming the company allegedly used around 12 million images to train the generator without a license. There have been other similar lawsuits from artists who are filing under copyright infringement. This is a great angle to take, but it only just touches the surface of the messy iceberg of all the ways AI can destroy the internet, the center of our first-world economy and culture for better or worse.
Can journalism take a similar stance? Can we hold generative AI companies and others who would like to pay for it accountable? That would be nice. It’s certainly possible in a class-action sense or if some of these corporations decide to do the right thing, but it feels unlikely, and not just because you need resources to do so. Journalism as an industry is too disparate. We’re in too many countries with different laws and ways of life, and these days, very few of us are a part of any journalism unions that could theoretically perform some kind of walkout. (It’s quite nice for these companies that so many of us have been laid off that we can’t band together in solidarity so easily. That’s probably a coincidence though.)
The definition of “freelancer” and what they’re allowed to do also varies from state to state, which is why a lot of unions or similar groups are local. The NY Freelancers Union is not a union, but it advocates for them and provides resources in a way that is similar to one. This and similar groups can’t register with the NLRB nor can they bargain under U.S. law, but they can be there in the event you need them.
There is a chance this could improve over the next decade. In 2022, Gallup found that 71% of Americans support labor unions, which is the highest it’s been in decades. As I’ve mentioned here before, there have been quite a few walkouts and union talks over the past couple of years, and they’ve resulted in some nice gains for those workers in terms of pay and benefits. However, these are only for a small percentage of people who classify themselves as journalists. I’m excited every time a media union succeeds, but there’s always a pang of jealousy because it doesn’t apply to many of us.
But it’s something. Strike Summer has put the thought of a journalism strike into my brain, and even though it’s, frankly, impossible, I do think it’s imperative that we support and stand in solidarity with the other striking groups, especially since we’re all working for the same things. Generative AI is impacting so many industries for the worse, and since the rich show no signs of throwing those AI plans away, we have to work against that however we can. I don’t have any answers for this myself — my expertise is in journalism and being angry about it, not on unionizing — but we can spread the word, talk about how harmful it’s been, teach people about the realities of generative AI, contact local officials in regards to potential legislation. That last point feels especially important considering so many of our aging politicians don’t know how to check their email, let alone understand the subtleties of emerging technology.
Until we figure out what to do as journalists, keep reading, keep educating. And regardless of what your boss wants you to do, maybe leave that new, fancy AI tool alone.
It’s been a while since my last post. I’m figuring things out on my end in terms of an alternative newsletter platform (Substack is… lacking in some areas, to say the least) and new way to do what I’d like to do. So stay tuned. In the meantime, if you’re so inclined, leave feedback in the comments below.