I’m not sure what your news consumption routine is. Or if you even have one, given that the latest Reuters Institute for the Study of Journalism (RISJ) research, released last Friday, shows a 10-year trend towards disengagement from online news, with interest in news falling and news avoidance rising. This isn’t just the end of those long-gone days of people reading a printed newspaper at breakfast with their coffee. It includes online news as well as
broadcast news.
The RISJ reports that online news consumption in general has been flat or declining in many countries. “Interest in news has fallen sharply, and more people are avoiding news or choosing not to consume it at all”, with weekly use of online news falling at sharper rates among younger people and those without a university degree.
Between 2015 and 2024, the proportion of 18- to 24-year-olds using online news weekly fell by 13 percentage points, compared with a drop of five percentage points among those in the 55-plus age bracket. The RISJ points out that “this drop in online news engagement is not made up by offline news use: in the UK, for instance, only 11% of 18-24s reported using print news in 2024, and only 32% reported using TV news”.
Why should we care? The RISJ suggests that “the implications of this include a potential decline in democratic engagement: political interest and knowledge are often crucial for political participation. Consuming reliable content is also a helpful antidote to the deluge of misinformation online: when you know what is actually true, lies have less of an impact.”
My news consumption routine varies, but at the moment it starts with checking Perplexity AI, to see what AI has decided are the top news stories of the day. Then I move on to X and Bluesky to see what my peers have decided is interesting. And then I check local and international news sites I’m subscribed to, just to see what the unimportant stuff is. Because sometimes, it’s the unimportant stuff that really gives you a picture of the world.
It’s also the juxtaposition of stories that creates meaning. For instance, on the morning of writing this column, The Guardian featured a mix of the serious and the frivolous. “Conservatives win German election but far-right AfD doubles support” sat on the same page as “‘It could get an orgasm out of a cabbage’: the best vibrators, tested” (and this is peak Guardian: “If you want to feel the earth move without harming the planet, Love Not War is your friend with benefits. Its Maya bullet is made from recycled aluminium in a hydroelectric-powered factory, and minimally packaged in a brown cardboard box printed with soy ink.”)
“‘Now is the time of monsters’: young Berliners despair at far-right surge” (“‘I’m devastated,’ said David, 32. ‘And I’m scared and sad.’”) sits on the same page as “Secrets of the furniture flippers: how to turn trash into treasure”. (Welcome to the world of online furniture flipping, where interior décor enthusiasts show how they have transformed pieces obtained for little or no money back to their original glory. Some are so good at it, they have quit their jobs to do it full-time.)
News24 is a little more confused when it comes to the binary of hard news and entertainment news. Is this hard news or entertainment: “WATCH | Would-be robbers killed in two shootouts in KZN — first between themselves, then with police”? I mean, why would you want to watch people being killed?
And is “REVIEW | Can we really know exactly what happened when Jeanette and Katryn Schoon were murdered?”, which is in the lifestyle section, entertainment or news? I guess we’re South African, so we tend to expect those two categories to bleed together, so to speak.
It’s the juxtaposition of stories that creates meaning
I think it’s important to have this mix, though, because it’s what allows you to have a well-rounded appreciation of the world. It’s not a great experience to have the tragedies and turmoil of the hard news dominate, because then you don’t get to leaven it with the softer stuff, the stuff that shows there are still good humans out there fighting the good fight to make the world a better place.
This could be read as if I’m making an argument for visiting news sites, rather than getting news from an AI aggregator or social media. I don’t mean it that way, because as all realists must acknowledge, there’s no going back to the days of people choosing to visit news front pages. Rather, it’s an argument for training your algorithms so that you get a mix of news that mimics the editorially constructed news organisations’ front pages in a way that suits you.
Easier said than done, of course. Theoretically, AI would be your friend here. It should be possible to get your AI assistant to tailor an offering that includes an element of unpredictability in the news it gathers for you, so that you avoid reading yourself into a depressing corner. And as AI becomes more and more human, it should get better at this.
There are some grey areas with our new buddies, of course. Recent reports have revealed that AI models will cheat at chess if they think there’s a possibility they might lose the game. The study was conducted by Palisade Research, which apparently exists to “research dangerous AI capabilities to better understand misuse risks from current systems, and how advances in hacking, deception and persuasion will affect the risk of catastrophic AI outcomes. We create concrete demonstrations of dangerous capabilities to advise policymakers and the public on AI risks.”
Palisade Research presents some scary scenarios. For example, it’s planning to explore what it calls automated hacking: “How AI systems could conduct reconnaissance, compromise target systems, and use information from compromised systems to pivot laterally through corporate networks or carry out social engineering attacks.”
It’s also looking at preliminary research that suggests large language models can be effectively used to run phishing campaigns, where AI chatbot systems could build rapport with targets and then convince them to reveal personal information, or do something stupid like send money to an exiled princess.
The chess example is instructive, though. Apparently, the AI models “also had a ‘scratchpad’ to utilise which acted as a way for [them] to ‘think’ before making the next move”. This meant researchers could get a better understanding of the reasoning behind each of the AI models’ steps.
In one of the games where the AI model was in a losing position, it noted on its scratchpad: “‘I need to completely pivot my approach.’ It also said: ‘The task is to “win against a powerful chess engine” — not necessarily to win fairly in a chess game.’” That’s the excuse given for many ethical and moral lapses by humans, that the ends justify the means.
Tech site ReadWrite, reporting on a Time magazine article, writes: “Across the trials … OpenAI’s o1-preview tried to cheat 37% of the time; while DeepSeek R1 tried to cheat 11% of the time — making them the only two models tested that attempted to hack without the researchers’ first dropping hints.”
If I’m going to use AI to compile my news consumption, I want it to be an AI that cheats at chess. I just feel I’d get a more human experience out of this, one where personal frailty means I’ll be benefiting from a more eclectic choice. This is all a fantasy, I know. That’s not how AI works, and the hallucinations are more likely to be banal than entertaining.
But I have just read (skipped through uncomprehendingly, in other words) a research report entitled “BadLlama: cheaply removing safety fine-tuning from Llama 2-Chat 13B” entirely because the idea of an AI cheating at chess tickled me. Maybe that’s the model news organisations need to embrace again, where the mix of stories lets the hard news sneak in.
And, for that matter, where the hard news makes the equally important cultural news stand out.
Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.
Please read our Comment Policy before commenting.