Welcome back to the Baekdal Plus newsletter. Today, I have a mix of things, but the most important is a look at the 85% market, followed by a note about AI vs journalism, and seasonal climate focus.
One of the things we know is that only about 15% of the public chooses to pay for news. This number obviously varies a bit. In Norway and Sweden, it's very high, whereas the rest of the world is seriously struggling to get people to pay.
This leaves us with a question. If only about 15% of the market chooses to pay for news, what does that mean? And more important, how do we then do something differently so that the remaining 85% could be convinced to also pay?
Obviously, we can't just create more news like we have always done because this is exactly what the 85% is saying that they don't think is worth paying for ... so what do we do?
Well, in my latest Plus report, I talk about this problem, why it exists, what elements dominate the trends, and what you should do about it.
So take a look at: "What if a newspaper was something completely different?"
Over the past couple of months, there have been a number of stories related to AI that have dominated my media feeds. One was a story about a Google employee who thought their chat AI had become sentient (which obviously wasn't true), and the other was a ton of articles about the new AI tools for generating images, text, and a lot of other things.
Every time we see stories like this, in the media we have a weird reaction to it. On one hand, we report about it as if it's magic, and on the other hand, we express outrage about the implications of it.
As a media analyst, however, I think we need to have very different discussions about this. Specifically, we need to talk about what AI means from a journalistic perspective.
Last month, Max Roser, from Our World in Data, asked:
I'm curious how you think about the limits of artificial intelligence. What is a task that you could ask a human to do right now which you think an AI won't be able to do within this century?
Now, the end of this century is a long way away from an AI perspective, but let's talk about what AIs are currently good at.
The first thing AIs are really good at is when you have a defined input that you can analyze, convert, or manipulate. For instance, if you have a data set about something, AIs can analyze this much quicker and often more accurately than we humans. So, things like detecting cancer cells in a defined type of picture, or being able to detect objects and classify them in videos.
They are not perfect, but already the AIs outperform us.
We also see this in journalism, where several newspapers are now using AI tools to automatically publish articles. This can be automated real estate updates, weather reports, sports news, etc.
But in every one of these cases, we have a defined and consistent source of data, and the AI is turning this data into something that is easier and more enjoyable to read for humans. I mean, as publishers, we are not in the Excel business. We are writers.
This is one area where AIs are really good already, and can be a great help for publishers.
The second thing AIs are good at is to 'make shit up'. There are many examples of this. One example might be how an AI is able to take just a few pictures and turn it into a 360° video, allowing you to spin around a scene.
Or we have tools like Dall-e 2, where you can simply tell it to create an image of "an astronaut riding a horse in a photorealistic style" ... and it will do so, in seconds. It will give you several variations to choose from.
These are really impressive, but the problem with this form of AI is that it's all based on lying to you. None of this is real. There were no astronauts on a horse. It made it up.
We see the same thing with so many chat systems. Yes, we know how chat systems are so good that they even fool Google employees into thinking they're sentient, but that too was designed to lie.
Here is a simple example for you. Go to one of these AI systems, and simply ask it what color it likes. It will most likely answer with some color, and then ask it why it picked that color.
Here is an example from a somewhat simple AI chat system.
First of all, it's really impressive that an AI can have a conversation like this, and if I had merely posted this picture on Twitter, without any explanation, many people would probably be fooled into thinking that Eliza is a human.
But the problem here is that it does this by lying to you. As an AI, it lives as a software app inside a server, so it was never 'little', nor did it go swimming. That's it telling me a lie so as to trick me into thinking that it's human.
And there are many tools like this out there. Another one is OpenAI's GPT-3. It has the capability to write very long text and the latest versions are so good at it that we humans cannot tell that it was written by an AI.
But it too is lying to you.
Mind you, there is nothing wrong with lying machines. We humans spend an enormous amount of time each day consuming lies, and most of the time it's quite enjoyable.
For instance, if you read "The Hobbit", none of the things in it are real. It's all something that J. R. R. Tolkien made up. Same with all the "CSI" TV shows. It's all made up. It looks real, but it's not. Nor did anything in Star Wars really happen.
So, as humans, we have no problem with spending time with lies. But think about this from a journalistic perspective.
As a journalist, you would never watch an episode of CSI, and then think that was real, and report it in the news like this:
Sure, it looks real, but as journalists we understand the difference between when something is just made up (fiction), and when something is based on reality (non-fiction).
But the problem that I see as a media analyst, is that we are not doing this when it comes to reporting about AIs. I come across stories almost every week, where the reporter hasn't made the distinction between whether the AI was designed to work with fiction, or whether they use actual data (non-fiction).
This is really important, and the example from above about CSI illustrates why. Fiction-based AIs cannot be used as a news story. It's not something we as journalists should focus on.
But, you may ask, how do we do this since we can't tell the difference? Well, you can, because you just need to ask what it was designed to do. Was it designed to 'make shit up', in which case it's the work of fiction. Or is it designed to analyze a defined data source, in which case it's the work of non-fiction.
So when we look at:
In other words, it doesn't matter how good these tools are, or how they are able to fool us humans. They were all designed to make shit up, and so none of that matters from the perspective of covering news, the work we do as journalists, or even the future of journalism.
This is what I do every time I hear about something related to AI. I ask, is it based on making things up, or is it designed to reflect on real world data? If it is the former, then I don't care. I might get amused by it from an entertainment perspective, but I don't care about it from a journalistic perspective.
Finally, I just want to point out a few things related to climate coverage. As a media analyst, this is one of my active focus areas because it's something that we have to get right as the press.
Over the past several months, I have noticed five things that keep repeating, and it's something we should think about making an effort to change.
Let me list what these are:
The first problem is not from me, but instead from Wolfgang Blau, who tweeted:
Climate journalism in the Northern hemisphere still has 'seasons', which it shouldn't. 'Oh look, it's hot again. We should write about it'.
This is such an important point, and something we see quite clearly. We talk about climate when the weather gets warmer, and that's obviously more of a concern (or so it seems) in the summer than at other times of year.
But this is the wrong climate focus. Climate change doesn't stop just when summer ends, and more important, it's not just about the temperature. In fact, there are so many other factors that are just as important (if not more) that aren't about how hot it is for a few weeks during the summer.
So, we need to change this narrative. It's not helping anything.
The second thing I often see is that climate action is very often reported about as a negative. We focus on the short-term cost of changing things, and we interview and feature the heavy polluting companies who now face rising risks and costs. On top of that, I often see articles saying that something is a 'win' for the petrol car owners.
Most recently there was a story about how people with hýbrid cars wouldn't face rising costs, which was reported as being good for them. In other words, many articles celebrate the status quo.
This is a terrible way to cover climate change, and this narrative encourages people to not act. What we should be doing is to talk about climate action in terms of the solutions and progress that it creates.
For instance, how do you talk about windmills? Do you focus on the cost of building them, or the savings in the future? I mean, right now we have a crippling level of inflation, more windmills could protect us from this in the future.
What about CO₂? Why do we need to focus on that? To stop climate change? Well, sure ... that's one important aspect of it. But it's also to help create cleaner air for us in the future. High levels of CO₂ impairs our cognitive abilities, so we are not just doing this for the climate. We are doing it for ourselves.
Same with fossil fuel, coal power plants, etc. ... Why do we stop those? Sure, again, to stop climate change, but also to reduce the already very high level of air pollution ... which is killing thousands of people every year.
I could give so many other examples, but the point is that climate action is about creating a better future. A future with better air quality, better cities, more healthy lifestyles, etc.
This is what we should focus our narratives on so that people feel a desire for action to happen.
Another problem I see very often, and that many other people have also pointed out is how the newspaper picture desk is not really up to speed about climate change.
Because it has been so hot during this summer, with record heat waves across Europe, we see a ton of articles about it ... all feature happy people enjoying a relaxing day at the beach eating ice cream.
It's not the right message to send when the temperatures are so high that people are dying from it. So, get your picture desk up to speed on climate action.
We also have a problem with the 'deadline'. Changing the climate is not an easy task, and revamping the energy infrastructure of an entire country is really difficult. So, when politicians talk about climate action, they talk about it as a deadline. They say that by 2030 this should happen, and then by 2040 this other should happen, and then by 2050 (long after all the politicians are gone from office), something else should happen.
The problem with this focus is that it makes the audience passive spectators, and it presents the problem as something that we have to wait for.
This focus might be fine for other news stories, but not climate change. Climate action is instead like an apple tree. Imagine that by 2030, you want to be able to eat your own apples, from your garden ... okay, so when do you plant the tree?
The answer is obvious. You plant it today. The earlier you act, the more apples you will have by 2030. This is how you should think about and cover climate action. The more efficient and determined we are acting today, the better the outcome later on.
So, as a media analyst, I don't care about 2030, 2040, or 2050. I care about what actions are taken today. But so often when I read about climate action in the press, I don't feel this sense of present focus.
Instead, I see a story about how Germany is now restarting coal power plants, because, from the perspective of the press, this is something that is needed right now (because of the situation with Russia), while climate action is 'later'.
That's not how this works. Climate change is NOW!
Finally, we have the 'if' problem. This is something that we still see in many newspapers that, when we talk about climate change, we still talk about 'if it will happen'.
I'm not going to talk too much about this other than to say that you should delete the word 'if' from your vocabulary. There is no 'if' when it comes to climate action. It's not a question, nor is it a debate. We have overwhelming data and evidence available to us.
So, stop using 'if' as part of your climate coverage. And when you interview an expert who uses the word 'if', don't use that interview because that expert clearly isn't the right person to talk to.
These are five things we should focus on changing in our newsrooms.
If you want to know more about climate coverage, I wrote a much more detailed Plus article about that last year called:
And also, don't forget this week's Plus report about how to capture the 85% of the market that currently don't want to pay for news.
Also, remember that while this newsletter is free for anyone to read, it's paid for by my subscribers to Baekdal Plus. So if you want to support this type of analysis and advice, subscribe to Baekdal Plus, which will also give you access to all my Plus reports (more than 300), and all the new ones (about 25 reports per year).
A look at the trend of brand+publisher, and the future for epaper
Asking an AI to do media analyst, and what does it mean when social becomes content focused?
It's tempting to just take a picture of your desk, but be mindful of what it might reveal
A guide to AI for publishers, the end of a million views, and what read metric is best?
Depression is impacting all level of news, from the journalists, the audience, to the businesses.
Founder, media analyst, author, and publisher. Follow on Twitter
"Thomas Baekdal is one of Scandinavia's most sought-after experts in the digitization of media companies. He has made himself known for his analysis of how digitization has changed the way we consume media."
Swedish business magazine, Resumé