The mystery of Gen Z's time online
Plus Apple enters the open source AI model game
Over at the FT, John Burn-Murdoch has produced a fascinating article based on some data about the opinions held by Gen Z, those aged under 30:
Gen Z is two generations, not one. In countries on every continent, an ideological gap has opened up between young men and women. Tens of millions of people who occupy the same cities, workplaces, classrooms and even homes no longer see eye-to-eye. In the US, Gallup data shows that after decades where the sexes were each spread roughly equally across liberal and conservative world views, women aged 18 to 30 are now 30 percentage points more liberal than their male contemporaries. That gap took just six years to open up. Germany also now shows a 30-point gap between increasingly conservative young men and progressive female contemporaries, and in the UK the gap is 25 points. In Poland last year, almost half of men aged 18-21 backed the hard-right Confederation party, compared to just a sixth of young women of the same age.
There are some remarkable graphs (Burn-Murdoch really made his name during the pandemic with his amazing graphs, and he’s continued that trend), plus this:
Outside the west, there are even more stark divisions. In South Korea there is now a yawning chasm between young men and women, and it’s a similar situation in China. In Africa, Tunisia shows the same pattern. Notably, in every country this dramatic split is either exclusive to the younger generation or far more pronounced there than among men and women in their thirties and upwards.
Alice Evans, one of the researchers on the topic, has written about this too: happily she’s on Substack, and (unlike the FT) her article on this topic isn’t paywalled.1
What’s driving this split? She sums up the causes thus:
Gendered ideological polarisation seems encouraged by:
Economic resentment
Social media filter bubbles
Cultural entrepreneurs
Cultural liberalisation, encouraging people to speak out.
Evans goes through the other three in order, but concludes that they aren’t the biggest problem (though they are a problem). But you know what she thinks is a problem? Social media filter bubbles.
The big, structural shift that coincides with the growing gender divide is technology. Any teen with a smart phone can play online games, watch comedy, browse social media or listen to podcasts. Whatever appeals, there’s unlimited amusement.
News reporting become more negative: highlighting terrible catastrophes. Web users are also hyper-connected, alert to horror stories worldwide. After a terror attack anywhere in Europe, German Twitter users show affinity with the far right.
One of the most interesting things about researching Social Warming was looking at the data from the PISA study, which gives tests on school subjects to schoolchildren from OECD countries, but also over a number of years broadened its questions to happiness, and since 2012 about internet access, and similar topics.
Here’s an interesting point that emerged from the PISA data, which I wrote in a chapter about social media and children, which regrettably had to be cut from the final book:
In 2003, South Korea led the world for the percentage of students with an internet connection at home, followed by Sweden, Canada, the Netherlands, Norway, Australia and Denmark. In eight of the countries, more than half of children said they’d been using a computer for at least five years, and not just to play games.
Also:
The PISA study in 2015 found that schoolchildren really liked social networks: 84% agreed with the statement “it is very useful to have social networks on the internet”. And they certainly used them. On average, across OECD countries, 91% of students had access to a smartphone at home—well ahead of those who could get hold of a laptop (74%), desktop (60%) or tablet computer (55%). Children had started using the devices almost as soon as they could read: on average across the OECD almost one in five had accessed the internet by their sixth birthday, and nearly two-thirds before the age of 10.
This is the Gen Z generation: eight years on they’re out of school, and they’re using their phones to live their life. The division that they find there would have been set out early on by the algorithms that they’ve been exposed to for years.
Yet the irony, given the results found by Evans, is that girls are far heavier users of social media, on average, than boys. For boys, their use of “computers” is often not spent so much on phones and/or social media, but on games consoles. Girls, in comparison, use social media far more heavily.
This points to a surprising possibility about social media use: perhaps longer social media use is what is leading to the more progressive positions of the young women in Gen Z, and the young men, with less experience of what the algorithm does, are vulnerable to being dragged to one extreme or another. In that reading, the polarisation is reduced in the women, and increased in the men. It seems like we want children and teenagers to spend more time on social media so that they won’t become so polarised.
Except.. it’s never quite so simple. Time spent on social media (and other systems) tends to follow an inverted bathtub shape: there’s a small number who spent next to no time on them, and then plenty who spend a reasonable amount of time, and a tiny number who spend a great deal of time. The first group tends to say they’re not happy about the amount of time they (don’t get to) spend on social networks; it turns out they want to spend more. The second, much bigger, group is perfectly happy. The third group, spending a lot of time online, is less happy. What isn’t clear is whether they’re unhappy because of the length of time they spend online and what they see there, or if they’re already unhappy and so spend time in the virtual space.
Does that give us a solution for the split between male and female in Gen Z, which shows no obvious sign of moderating? The easy answer is to say “ban smartphones”, but do we also want to lose the benefits that we’re seeing for women? Evans’s research points to how it can be done: people just need to mix more in real life.
In Catalonia today, feminism is a common topic of conversation. Young women are publicly criticising inequalities; some are also educating their male friends. Santiago (who’s just finished school) shared examples of his friends resisting machismo.
After being jilted, one guy said women are whores (puta). She replied “No, a woman may decide who they wish to go out with”.
Groups of guys can get rowdy, especially when drinking and watching football. Santiago’s female friends complained, they wanted to leave. He learnt that aggression made them uncomfortable and thereafter he became more sensitive.
Kissing on the cheek is a traditional Spanish greeting. But Santiago’s female friends find this too intrusive. They prefer to shake hands with strangers. By speaking out and supporting each other, young Catalonians are creating a public sphere in which women feel more comfortable.
Seeing their discomfort and defiance, keen to preserve their friendship, Santiago listens and learns. He actually feels more comfortable with women, they don’t pressure him to be macho.
Sounds simple, doesn’t it? Let’s hope that over time this is going to be how things work. Because if not then we’re all in a lot of trouble.
Glimpses of the AI tsunami
(Of the what? Read here. And then the update.)
• Quora is turning into a junkyard, because it has added an AI answer generator to what used to be a reliable-ish question/answer site. (Sturgeon’s Law applies even before the AI was set loose.)
• AI image generators are starting to fill up Pinterest and Instagram, particularly for house pr0n. The big question is whether it can really come up with a new interior design that you’d like.
• Google Maps is getting chatty with a new update from its LLM system, which has also been released separately as Gemini.
• Apple (yes them!) with the University of California at Santa Barbara has released an open source AI system which can do image editing from written commands. It’s called MGIE. I can’t quite figure out why you’d want to do image editing with written commands unless you had a disability or found yourself in a peculiar situation that required it. But more generally, Apple getting into open source AI models has to be seen in the strategic framework: in business, make your rivals’ valuable things available for the lowest price possible. If LLMs are valuable to Google and Microsoft, making open source ones available for free is a good way to undermine their expected future revenues.
• You can buy Social Warming in paperback, hardback or ebook via One World Publications, or order it through your friendly local bookstore. Or listen to me read it on Audible.
You could also sign up for The Overspill, a daily list of links with short extracts and brief commentary on things I find interesting in tech, science, medicine, politics and any other topic that takes my fancy.
• Back next week! Or leave a comment here, or in the Substack chat, or Substack Notes, or write it in a letter and put it in a bottle so that The Police write a song about it after it falls through a wormhole and goes back in time.
She also has the graphs from Burn-Murdoch’s piece, but I’ll leave them there because he quoted her research, so I think there’s a quid pro quo between them.