Assigned Media Commits to Staying AI-Free
As fancy chatbots wash over the media landscape, here’s what you can expect from your favorite transgender news site.
by Evan Urquhart
Now cryptocurrencies have gone the way of the metaverse, the tech world’s new hotness is fancy chatbots. Programs such as ChatGPT and its competitors are often misleadingly referred to as artificial intelligence, especially by their boosters. This misnomer has led to widespread misunderstanding of the function and capacities of these programs, which do not contain any intelligence in the sense of having an awareness of the real world or facts within it. They are computer programs that use statistical models to predict the sorts of things a human would write, quickly producing such a fascimile. This notably introduces errors when a chatbot is asked to produce text when they don’t have pre-existing text about that topic available to remix: In those situations they produce something that sounds true, but is entirely fabricated. Crucially, they do this without knowing that they’ve provided false information.
Already fancy chatbots have begun to be used for their primary purpose: To allow complete assholes to lay off human writers. It remains to be seen whether the output of chatbots will be high quality enough to accomplish this mass layoff plan successfully, but the intent seems to be to start by firing writers with the rationale that their core job, writing, can now be automated. Next, because the core job of reporters is actually reporting, outlets that are concerned with accuracy will create a new set of much more precarious positions (with lower wages and less stability). These new writers, who probably won’t be called reporters, will produce reporting for the chatbots to write from and also check the chatbots work (meaninh they’ll do the same research as a human reporter would otherwise have had to, but at a lower pay rate) to ensure accuracy.
Apart from driving even more the layoffs in a struggling industry, the main promise of chatbots in news seems to one of be producing low-quality, error laden text quickly. Assigned Media is written by a single trans guy and manages to create one or more posts every weekday, and two newsletters weekly. As such, there just doesn’t seem to be much need to produce even more content, even faster. Additionally, we’ve found the errors Evan already makes (most often spelling or gramma) to be more than sufficient for our needs at present.
If the promise of chatbot technology is to automate the easiest, quickest, most enjoyable parts of writers’ work it seems it will be no great sacrifice if we commit to soldiering on without it.
However, that is not the only way we expect chatbot technology will impact Assigned Media.
Our core function is to monitor, analyze, and fact-check anti-trans propaganda, particularly on the US right. As such, we know well that right-wing news is run by people who put business principles above moral ones, and such outlets already produce a high volume of low-quality, inaccurate, and misleading content. They also continually spawn new outlets with dubious provenance, and generally do as much to muck up the information ecosystem as they are humanly capable of. Just imagine what will become possible if they are no longer bound by human limitations!
The outlets we monitor seem highly likely to become early adopters in the rush to replace low-quality human content with low-quality automated content. For the least accurate, poorest quality sites we expect there will be little actual difference. We judge that the greatest risk will be with right-wing outlets that present themselves as relatively reputable, outlets such as Fox News, the Federalist, the National Review, tabloids like the Washington Examiner and the New York Post, and activist “think-tank” publications such as the Daily Signal and City Journal. These are publications where, most of the time, the information published, while misleading or removed from context, has not been made uou of whole cloth. It seems possible that this state of affairs could degrade in unpredictable ways due to a rush to replace humans with chatbots, either because the sites they trawl for stories could undergo a complete collapse in their relationship to real world events, or because these outlets also start relying on chatbots for their content.
Mainstream outlets are also owned by the layoff-hungry sociopaths of the business world. While thetop tier of the mainstream seem the least likely to be impacted, 2nd and 3rd tier publications are very much at risk, and there’s also a high risk of quality collapse in smaller local news sources. For Assigned’s purposes this may make it harder to know for sure if some obscure local story that has right-wingers frothing at the mouth was accurately and independently reported by a human at a local mainstream outlet, or if we’re seeing a chatbot remixing the right’s fake content with a more mainstream style.
This is, not to put too fine a point on it, a bit of a conundrum for a news site dedicated to monitoring and reporting accurately on anti-trans narratives in media.
Right now, the commitment we can make is that we’re very aware of the risks of chatbot-takeover of news organizations, and watching this all closely. We will work to understand which news outlets are or will be using chatbots, what the policies of different outlets are when it comes to artificially-generated text, how carefully such text is being vetted by their newly degraded human workforces, and so on. If we’re aware that an outlet we’re reporting on uses artificially-generated text, we’ll spotlight that. We’ll make an effort to investigate the bylines for stories we’re covering and, if we can, verify that these names represent actual living humans. As the new hotness of chatbot technology matures, we may check in again, and give readers more information on how chatbot tech is interacting with anti-trans propaganda.
We are, frankly, a little pessimistic about this interaction. Legitimate-sounding mega-viral anti-trans stories that will be unusually hard to figure out the truth behind seem likely, perhaps inevitable. Our commitment will be to be transparent about our reporting process and what we do and do not know, and as always remain open to updating or correcting a story when necessary.
So-called AI may be as bit a bust as crypto, or it may change the news landscape completely. If it’s the latter, it will almost certainly not be for the better. Our experience suggests that right wing media will be on the cutting edge of developments in ever cheaper, lower-quality written content. Assigned will keep watching, and reporting, whatever comes of it.