Journalists Are Digging Their Own Graves
The debate of how AI can be best used in journalism is completely beside the point

Düsseldorf, 9 August 2023 Filed under: Technology
The hottest thing in journalism right now is to debate how AI — a marketing term which most journalists use to abstract more meaningful terms like machine learning — can be used to rescue the profession. I’ve been hearing about this everywhere lately, from predictable sources like the Silicon Valley hype press to unlikely candidates like official publications by Germany’s biggest journalism union DJV. This concept so incredibly misguided that following the idea to its logical conclusion would actually mean the death of journalism as we know it. Naturally, the proponents of using AI in journalism do not understand this at all.
Anyone with half an ounce of common sense to their name can see the shape of the crisis journalism is having — except most journalists. They don’t understand that their proposed solution is actually the biggest problem journalism is facing right now, because they don’t understand the crisis in the first place. These journalists have been co-opted by media company executives, thinking that the problems of publishers and network owners are their own problems. Which is ludicrous.
Out of the Frying Pan, into the Fire
Media company bosses noticed that journalism — as a business proposition, mind you — was in trouble when advertising revenue started dropping due to competition from internet services. They realised that they could not compensate for these losses by increasing subscriptions to their content, so they started to look for another way to keep the business profitable. Which was to lay off good people — brilliant writers, scathing critics, ruthless analysts — and replace them with fewer people, who, by and large, were willing to work more hours for less money. Most of these new people weren’t as good as the old people. Some of them were as good, but they weren’t given the resources (in time, money and backing by their editors in chief) to produce the same quality of work. This saved some money and compensated for the falling income somewhat.
These executives still think that this crisis of revenue is what is ailing journalism today. And they, somehow, have convinced many journalists of this. In reality, this business crisis bred the actual catastrophe that is threatening the very existence of journalism right now: a crisis of trust.
A Very Simple Proposition
You see, at its core, journalism is defined by a very simple proposition. The audience hires a journalist to explain the things to them that the audience needs to know. It pays the journalist to find out what things should be relevant to them, to go and learn about these things and then explain them in a way that is convenient to the audience. Because the members of the audience don’t have time to research these things, by and large having their own jobs to worry about, they commission journalists to spent this time for them. This requires trust. If the audience does not trust the journalist, he or she is of no use to them.
Because the media executives ran into some trouble on their side of this business — organising a way for the audience to pay the journalists — they degraded the quality of the journalists’ work. Because they still wanted to make a lot of money. But this has led to the very widespread phenomenon of journalists losing the trust of their audience. Because, in all fairness, they by and large have been doing a shit job.
So now the executives’ solution is to actually do away with journalism altogether, hoping the public won’t notice. This isn’t hyperbole. Asking how one can use AI in journalism is akin to asking how one could make a plane fly without the principle of lift. You can’t. Where lift keeps a plane in the air, trust keeps a journalist their audience. You can’t trust a computer algorithm. That’s not how human minds work. At least not the ones who possess an iota of self-reflection and the ability to think critically about the answers they are given to the questions they are having.
So-called artificial intelligence isn’t intelligence. That’s marketing bullshit. What we call AI is a very powerful statistical analysis based on a lot of data that has been scraped from the internet. And as everyone knows — journalists have been crying it from the rooftops for years, after all — the internet is full of bullshit, some of it dangerous. You need a human to sort the crap from the good info. A statistical analysis can’t do that. Especially if it was trained on the bullshit in the first place. You need a human to tell Midjourney that humans have five fingers, one of it a thumb, and no other combination thereof. And you need a human to tell ChatGPT that the JavaScript you just asked it to write has a serious security flaw that makes it very dangerous to use. You need a human.
Without Trust, We Have Nothing
If audiences don’t need a human they can trust, they don’t need journalism at all. They can just ask Google what they want to know. If they don’t care if what they get is good, dependable information or dumb propaganda from someone with a vested interest, journalism is dead already. Its one and only chance to survive is to emphasise why you want a human working for you. To explain why statistical analyses can be dangerous and how they can at best fall prey to unseen biases and at worst be intentionally misused to misrepresent reality. Lies, damned lies, and statistics.
Yes, you can use machine learning to generate some art to decorate your article with (see above). Because that isn’t journalism — unless you’re using the image to illustrate a point — it’s advertising. It might be questionable when it comes to supporting the livelihood of artists — and it will certainly be of lesser quality than hiring an artist, because AI precisely isn’t intelligent — but it is not morally questionable. A piece of art is evaluated by the audience based on its inherent merit. It doesn’t really matter if a person or a machine or a combination of both made it. It doesn’t set out to prove a fact or sway people to an opinion, after all. But journalistic content does. It is by definition trying to be trustworthy. Replacing the mind of a human with a statistical analysis is misleading and corrupts the nature of the work — and its relationship with its audience. It is morally very questionable and people will react badly to it — as well they should.
The only way forward for journalists is to understand why they are having a crisis with their audience and to fix it by taking their audience seriously again. If journalists make the crisis worse by doing the opposite of what their audience wants, journalism disposes of its own usefulness. And in that case, journalists can hardly complain — they brought this fate on themselves.