In the beginning: an idea is born
Plus Facebook is turning text into video, and it's scary and thrilling
“Ice will melt, water will boil/ You and I/ Will shake off this mortal coil” — It’s Only Natural, Crowded House
Where do ideas come from? Where’s the point where you tie two or three disparate thoughts together and they coalesce to form a single cohesive thing? With Isaac Newton, we’re told the crystallising moment was an apple falling on his head. For Charles Goodyear, it was dropping some rubber mixture into a frying pan and seeing what happened. For Alexander Fleming, it was a petri dish left on a windowsill, where a fungus demonstrated its antibacterial properties on the culture growing there. For Einstein, it was going for a walk with a friend and chatting about a puzzle that had haunted him for nearly a decade.
I can’t claim to be anything like as exalted as those folks. We just know the origin stories of their ideas very well; they’re a sort of folk history. But I thought I might as well tell the story of how “social warming” struck me as being a phenomenon worth describing, and perhaps even testing against real-world examples.
The prompt was being asked to give a lecture at Cambridge University to its arm of the Philomathia foundation1 in November 2017 with the title (which I didn’t choose) of “Democracy and communications technology: which controls which?”. I had already been attached as a research fellow to the university’s Technology & Democracy group, run by John Naughton and David Runciman, where I’d been looking at the question of whether better internet access led to people becoming more, or less extreme in their views. A YouGov survey, carried out in April 2017 for this research, showed a strange dumbbell effect: given how strongly or weakly they opposed or supported various policies (rail nationalisation, assisted dying, the death penalty, banning abortion, unilaterally abandoning nuclear weapons, legalising cannabis), the strongest support or opposition came from those who spent the most time online—or the least time online. I think the latter could be called “the Daily Express effect”: that paper’s readership skews very old (typically aged 69; 83% are over 55). The older demographic doesn’t get exposed to many conflicting viewpoints online, because they tend not to be online.
Online all the time
The former effect, where the most extreme positions are held by those who are very online, seemed more explicable when you consider Cass Sunstein’s 1999 research about group behaviour (covered in Social Warming), which finds that if you get a group of people who agree on a topic together and let them discuss it, their collective view will shift towards the extreme version of that view; that is, the group becomes polarised. This seemed to me interesting, and salient: the internet on its own isn’t a radicalising force, but seeing lots of views you already agree with will tend to take you further. People who spent really long times online (the range on offer went from “several times a day” to “once a day” to “3-4 times a week” down to “less than once a week” and “never”) were as polarised as those who were hardly online. But those who spent middling amounts online? Not so polarised. Something was afoot.
In preparing the lecture, the obvious thing to do was to look for examples where either democracy affected what communications technology was available, or where communications technology affected democracy.
It didn’t take much thinking, in 2017, to come up with a repudiation for the first proposition. That made up my third slide:
QED.
So, on to the next proposition. Can communications technology affect democracy?
Communicate this
I’d been writing about smartphones and the explosive growth in their use, especially since 2010. I could reel off the examples. The Arab Spring of 2011, including Egypt’s attempt to forestall Hosni Mubarak’s overthrow by cutting off the internet, was an obvious one.
Then there were a couple closer to home, both in location and time: Brexit, and Trump’s election. For Brexit, the full details hadn’t yet begun to come out about Facebook’s role in the referendum (Carole Cadwalladr had begun writing her series of stories in May 2017, but there was much more to come). Trump’s seemed a lot easier to analyse.
Looking at the detail, I was amazed. He won the 2016 presidential election by getting the electoral vote in three formerly Democratic states: Wisconsin, Michigan, Pennsylvania. He won the votes in those states by utterly tiny margins: 10,704 votes in Michigan, 0.25% of the total votes cast; 22,748 votes in Wisconsin, 0.81% of the total votes cast; 44,292 in Pennsylvania, 0.75% of the votes cast.
In all, Trump won 77,000 votes more than Hillary Clinton in those three states, out of 13.2 million votes cast; a 0.6% difference. Six people in a thousand. But the electoral college system means that even if you only win by a tiny margin, you get all of the electoral votes for that state. A tiny difference can make a huge difference, rather like ice turning to water at its phase change at 0ºC, or from water to steam at 100ºC.
What we also knew was that Trump’s campaign had made very heavy use of Facebook advertising, unlike the Clinton campaign, which stuck with TV advertising; Clinton didn’t even visit the states because her team was so sure she’d win them.
What if, I wondered, those Facebook adverts had made some tiny difference. It wouldn’t have to be a big difference. Just enough.
In the same way, global warming had been on my mind for a long time. Looking around, everyone was talking about the importance of taking action on emissions. But people just weren’t. They’d take short trips in their car and then tut about Amazon logging. As a species, we’re amazingly bad at seeing what we do from any sort of helicopter view, and relentlessly selfish. Meanwhile, the glaciers drip into the sea, as tiny amounts of warming—it’s only 0.1ºC!—led to big differences.
Maybe the same thing was going on, I thought. Maybe social media is precipitating some sort of change in its users. Facebook was gigantic in the US in 2016, reaching rural parts that other media perhaps didn’t. (There was lots of media coverage of Trump voters pre- and post-election; an incredibly common thread was heavy social media use.) What if it was causing some sort of polarising effect on users, as I’d seen in the earlier research, and that was having wider effects on society?
Global warming; so obviously social warming. And so, the fifth slide in the presentation, which must have been the first use of the phrase:
And there you have it: as much of an explanation as I can muster for how I came up with the idea. Not as good as Newton’s, Fleming’s or Einstein’s, but we do what we can.
Glimpses of the AI tsunami
(Of the what? Read here.)
• Facebook has opened the kimono ever so slightly to show off a text-to-video system it calls Make-a-Video (not sure how much work went into the name). Give it a text prompt and it generates a small 5-second looping video:
“With just a few words or lines of text, Make-A-Video can bring imagination to life and create one-of-a-kind videos full of vivid colors [sic] and landscapes.”
Definitely true, as with the GIF, via Meta, below of “a teddy bear painting a portrait”. More details in James Vincent’s story at The Verge. This is amazing.
• OpenAI, those jolly people who gave us GPT-3 (the AI system which can generate entirely believable text from a short prompt: as that link says, it’s like a freshly hired intern who is well-read, opinionated and has a poor short-term memory), have now produced a speech-to-text system called Whisper. It’s hardly the first speech-to-text system (there’s one on your phone), but early reports say that it’s really good, dealing with mumbling, spotting names and capitalising them, all those desirable things. It presently works in English but has a big corpus of other languages that it can be used on.
There’s downloadable code, but requires some expertise to install; as with Stable Diffusion, expect it to be an app that can run locally within a month.
• Darth Vader’s voice (originally by James Earl Jones) is now being created by a speech-generating algorithm (from a team in Ukraine, as it happens):
What Respeecher could do better than anyone was re-create the unforgettably menacing way that Jones, now 91, sounded half a lifetime ago… “For a character such as Darth Vader, who might have 50 lines on a show, I might have a back-and-forth of almost over 10,000 files.”
• An “AI artist”—though he says we should use the term “synthetic media”—talked about his views of what it will and won’t do to work and jobs at The Atlantic.
• Missed this when it happened, but in June, Cosmopolitan magazine had an AI-generated cover (in “the AI issue”—can’t think how I didn’t see that; has my subscription expired?), and an accompanying article about working with Dall-E to do it:
if you type “koala riding motorcycle,” DALL-E draws on what it knows about koalas, motorcycles, and the concept of riding to put together a logical interpretation. This understanding of relationships can be keen and contextual: Type “Darth Vader on a Cosmopolitan magazine cover” and DALL-E doesn’t just cut and paste a photo of Darth; it dresses him in a gown and gives him hot-pink lipstick.
Can definitely recommend looking at the results from that one. (Mr V seems to be figuring heavily this week.) The whole article is excellent, in fact.
• At the Washington Post, Nitasha Tiku looks at some of the ethical concerns being raised by AI illustration systems. “It’s an adversarial environment, like all social media and chat systems and the internet,” according to David Holz, CEO of Midjourney.
• 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.
“Our mission is to promote human harmony and wellbeing through advocacy for, and support of, education and research – in particular, the exploration of new ideas in science, technology, the humanities and philosophy.”