In 2015, I wrote an article for my previous publication that said humans were intrinsically scared of the thought of a world run by machines, even though they shouldn’t be. That artificial intelligence was something we should not fear, but rather embrace.
If you can’t be bothered to read that, I don’t blame you. But to summarise, it basically was a piece telling people their jobs would be enhanced, not replaced. That AI was good, actually: It would help speed up mundane tasks, make life easier, etc.
At the time, I clearly never thought my job would be one of the first they would try and rid.
I say try, because it’s quite comical, the attempts that have been made.
That article I referred to just now, was run at ZDNet, which has now shuttered its Australian operation. ZDNet (still live in the U.S.) falls under the CNET Media Group, which, after an acquisition, is owned by a large U.S. firm called Red Ventures.
CNET, once the go-to place for everything consumer tech, is now an affiliate paradise. But that’s a whole other story for a whole other time. Earlier this year, CNET was caught publishing articles written by ChatGPT, many of which were full of serious inaccuracies. The publication had not told readers about the use of AI.
Those links take you to the U.S. Gizmodo, as I personally blocked them from landing on the Australia edition, mostly because they named individuals I called my friends and I wasn’t in the business of shitting on other publications. This is all important context for my rant to come, I swear. Anyway. CNET laid off 10 per cent of its staff weeks later, though the company said the move was unrelated to AI.
But it’s not just CNET – they were just first (to be caught out, anyway) and the first very close to my heart.
Gizmodo, part of G/O Media (boasting kick-ass brands in the U.S. like Kotaku, Jalopnik, AV Club, The Onion, etc) had its own controversy with publishing factually incorrect AI-written articles in July – Gizmodo Australia didn’t let those articles through to the AU site, either, obviously, but there was one about Star Wars on Gizmodo.com and over on Jalopnik (the site we syndicate car content from), they gave the Jalopnik Bot the opportunity to write two things in August, including this… slideshow.
Locally, we’ve got News Corp, and there’s also Bragg Media, which admirably decided to not be coy about it and tell readers they were going to be playing around with some AI. Other publications, however, have not been so open – with a handful of articles circulating Aussie media groups and being called out for blatant copyright and/or ChatGPT-esque sentence structure.
So anyway, here we are, in 2023, where AI definitely has its benefits. I used Bard just Sunday to help me build a ‘low-calorie vegan meal plan for a week’ because I am very sick of having to work that out for myself. But, I’m just not sure it has application in this industry. AI is only as good as the data that feeds it, and a lot of the data that feeds AI comes from articles about things. I’d love to see an AI watch Parliamentary hearings for a week, differentiate between speakers, add context of what happened previously in those industries, add context relevant for the specific site’s readers, and find the “gold”, the “lede”, the actual story and not just summarise a day’s worth of testimony. I’d also like to see a bot review a brand-new television or mobile phone to meet global embargo (meaning there is no information out there on the actual device, just rumours). The AI bot, no matter how advanced, would not be capable of that. They could, however, write a list of things to watch on Netflix if you don’t want opinion, you just want a list.
It’s a perfect tool for content farms and to put it simply, it sucks that a lot of people making decisions don’t see the difference between investigative journalism and churnalism when the end result, a page view, is what’s measured.
But there’s also the very real issue of AI bots being riddled with factual errors.
Even with human oversight, there’s a very real concern it’s dangerous to lean on a machine to produce accurate, unbiased facts. But that’s not the point. The point is reduce a line item on a budget (a journalist who spent four years at university attaining their degree) and retain the same output (article count), with the same performance (clicks), and sneak in a link to buy something.
If media organisations think AI is the way forward, reducing newsrooms to smaller and smaller amounts, relying on this innovative tech to take the lead – they’ll soon realise readers don’t read shit that’s regurgitated up by software just because they’ve been loyal readers in the past. But if they don’t realise this, dear readers, please, make your dislike known. Don’t read anything written by AI – they’ll soon realise the biggest mistake they made was getting rid of, and not investing in, the very people that made them so successful in the first place.