Cognition Dissemination: AI Isn’t Cutting It
Several stories have been written about the increasing usage of AI in the workplace, including on this very blog by yours truly. It’s a situation worth taking seriously. Since the advent of programs capable of writing copy and articles, ChatGPT chief among them, there’s been open mulling about whether they could fulfill the jobs of humans in offices (inside the home and outside) around the world. The technology has similarly been worth watching for artists. Just as there were warnings about automation subsuming blue collar jobs, the worry has now spread to white collar variants.
The developments are worth watching, and certain bosses have clearly been giddy about replacing workers who demand fair wages and benefits with robots with no such desires, even if they haven’t shown it too outwardly. Fortunately for the workers, this isn’t working out as well as expected for those which have tried it, and even the smaller number of publishers and companies which have sought to use AI for artists — the latter primarily outside the United States.
Previous examples regarding AI problems have manifested since the year’s start, one of the biggest including CNET using programs to provide daily financial news while making constant factual-sounding errors. But that was never going to stop other publications from giving it a shot. Take G/O Media, a company previously in the news for closing leftist politics blog Splinter (with the staffers openly criticizing the leadership style of the new company heads undoubtedly being part of that) and causing exoduses in other blogs like Deadspin, The AV Club, Jezebel, and The Root. G/O’s higher-ups sprung a surprise on the writing staffers at the end of June by informing them of an impending test for AI-written posts on certain blogs. The staffer reactions were not surprising, alongside the result.
AI posts left to their own literal devices have a habit of stringing together factual-sounding lies, and G/O’s experiment followed suit with its article listing Star Wars works in chronological order. (I’m not linking to it here.) Outside the dryly written text in the article bylined by “Gizmodo Bot,” itself a problem, serious mistakes were made. The Clone Wars and Rebels, two animated shows, are listed as taking place after The Rise of Skywalker, which is not true. Clone Wars fills in the gaps between Attack of the Clones and Revenge of the Sith, while Rebels occurs before Rogue One and A New Hope. The list also doesn’t include The Book of Boba Fett (alongside The Mandalorian), Obi-Wan Kenobi (not too long after Solo), and Andor (in tandem with Rebels).
A chronological Star Wars list couldn’t be a more basic assignment. If AI can’t create this by itself, it is not ready for primetime. The better idea would be to have a human edit the list after AI creates it, but that’s not a good idea either. The best idea, of course, would have been to merely hire a human to write it in the first place, though it’s not like such timelines haven’t already been assembled across the internet, some of which also include other media like the comics and video games.
This is also affecting voice acting, albeit to a lesser extent… for now. AI bros had the nerve to bully the great Erica Lindbeck off Twitter after she merely asked some of them to stop using her AI created voice for memes involving Persona 5’s Futaba Sakura and told fans to report them. Several voice actors have come to her aid and have mentioned their distaste for the AI trend, regardless of their usage for “hilarious” internet material. AI voicework isn’t up to snuff either, as it’s been unable to mimic emotional vocal ticks, but this issue is alone giving the actors pause for good reason. If they don’t stop it now, it will run amok.
AI work for writing, artwork, and voicework isn’t ready for primetime now, but doesn’t mean it won’t be in the future. Considering how quickly these developments have progressed, to the point that few outside big tech circles figured the technology would reach this point at this time last year, solid advancements could happen in the near future. The livelihoods of many could be significantly interrupted due to their jobs being subsumed by machines.
There could come a time when companies and bosses regret chasing the AI trend, with both them and the audience missing the human touch essential for this work. It’s very easy to tell when something’s written by Skynet or not, no matter how convincing it tries to sound. It has the potential to end up the same way as the “pivot to video” trend. But like that trend, several workers could use their livelihoods in the meantime, and there’s no guarantee all those jobs lost will return. If only modern society wasn’t too cynical for a Universal Basic Income trend to take off on a significant scale.
About The Author
Geoff
News Editor | Geoff is some nerd who writes for a blog and thinks he's knowledgeable. He write a lot about video games, but is capable of writing about other subjects, too. Twitter
