When basically adequate things turn bad

This week: how AI turned Quora into the fully automated luxury mansplaining machine! But first: the case for firing Ed Davey.

For half a year, half a decade ago, I wrote a weekly column about a man most people had never even heard of. Daniel Hannan MEP had spent so long writing ostensibly thoughtful but factually dubious columns attacking the European Union – pausing only occasionally to appear on US television to attack the NHS – that he’d made himself a bête noire for much of the liberal, pro-European left despite being nobody very much. “Hannan Fodder”, as I called the column, did surprisingly good things for both the New Statesman’s web traffic and my own Twitter following. It was fun. 

After a few months of this, though, I stopped. Partly because I was repeating myself; partly because I was bored. But largely, it must be said, because it had gradually dawned on me that the existence of a pro-European columnist, obsessed with Hannan specifically, was almost certainly doing far more to raise his standing with those he cared about than it was to reduce him in the eyes of everybody else. (Anyone who did not already hold a strong view was unlikely even to be reading.) This had not been the point of the exercise, and so I went back to ignoring him again.

In the same way, I’ve often wondered if a review so stinking it goes viral can actually be good for a book or a play – that the “alerting people to its existence” function might count for more than the “telling them it’s bad” bit. It’s not quite true to say that all publicity is good publicity. But there are situations in which no publicity is far, far worse. 

Which brings me, in a roundabout sort of way, to the Liberal Democrats. The party has long specialised in bizarre stunts: demolishing a symbolic blue wall with a tractor; unveiling a giant clock to call time on a struggling government; making a frankly horrifying puppet show concerning Jeremy Corbyn. This enthusiasm for stunts so cringe-worthy that you feel embarrassed merely to be sharing a planet with them is, with apologies to a significant minority of my readership, at least partly a result of the type of people who decide to join the Liberal Democrats. But there is method in this madness: as a third-party without either power or the prospect of getting any, it’s a constant battle to get any coverage at all. The clear calculation is that the positive effects of these stunts (reminding people you exist) counts for more, electorally, than the negative effects (reminding people you’re a bunch of dorks). 

All of which is why I think there’s an extremely strong case for sacking the party’s leader. Not despite the fact it’s an election year: because of it.

The difficulty of persuading AI art platforms to include Ed Davey in pictures, even when specifically prompted for him, feels like a telling problem in itself. Image: Craiyon.

A quick precis for readers outside Britain: the political news here these past two weeks has been dominated by a long-rumbling scandal concerning the Post Office, Fujitsu, a broken IT system and the biggest miscarriage of justice in British history. This has been sparked not by any parliamentary or judicial action, but by an ITV drama series starring Toby Jones. (In these times, you take what you can get.) Lib Dem leader Ed Davey played an extremely small part in the story by declining, while serving as a junior minister in the coalition government back in 2010, to meet Alan Bates, the campaigner and ex-subpostmaster whose name is memorialised in the title of that drama. “I do not believe a meeting would serve any useful purpose,” he said.

As it happens, Davey later did meet Bates. He blamed his initial refusal on the fact he was lied to by the Post Office’s management. And he is very far from the only politician to have failed to tackle this scandal while a minister: the procurement process for the Horizon accounting system which led to all this dates back literally 30 years. It would be unfair in the extreme to hold Davey responsible for this, while letting sitting government ministers off the hook.

But, as Rishi Sunak currently looks increasingly likely to howl into a microphone any day now, politics isn’t fair, and an untimely end to Davey’s career might be a smart move for his party. A change of leadership would cut the party’s last ties with the coalition years, allowing it to cast itself as in opposition to this terrible government, not a one-time enabler of it. It would allow the party to pitch itself as one which believes, unlike certain other parties one might mention, in taking responsibility for your mistakes. 

And it would remind people, in the run up to a general election, that the Liberal Democrats exist, and that they are not like the Tories. 

There are counter arguments. Some of them are bad: one activist suggested to me that it’d be unfair for Davey to resign while no Tory does, as if holding yourself to a higher standard than your opponents is the sort of thing which puts the voters off. Others, though, are more convincing. Any resignation statement that isn’t very carefully designed risks being clipped into something that’d make it look like this was all Davey’s own personal fault, which it very clearly wasn’t. And one Lib Dem friend points out that the party’s internal democracy is so tortuous that there’s a non-zero chance it might go into an election without a leader. The strategy is not risk free.

But polling last weekend suggested that this is the first thing Davey has done to really impinge on the consciousness of the nation, and it’s sent his personal rating plunging. My suspicion is, like bad reviews of a play or my column on Daniel Hannan, a graceful resignation from Davey, which calls out those Tories who’ve not had the guts to follow suit, would win the party more votes than ignoring the scandal would.

It’d be a whole lot less dorky than those puppets, too.¹

Fully automated luxury mansplaining

I have, in my time, thrown a lot of questions at the internet. Why is there a Napoleon I and Napoleon III, but no Napoleon II.² Why is Europe considered a continent when it is very clearly just a peninsula of Asia.³ Sometimes the answers to these questions end up in this newsletter as content; sometimes they end up coming out of my mouth in social situations, surprising and delighting my friends. 

At some point – I’m not sure when – such questions led me to sign up to Quora, “a platform to ask questions and connect with people who contribute unique insights and quality answers”, which, annoyingly, you need to log into to read. I don’t remember what question led me there, let alone the quality of the answer, but I’m assuming – given both my interests, and the sort of thing it’s been spamming me with ever since – it was somewhere in the intersection of history, national identity and international relations. At any rate, for several years now, the site has been emailing me other examples of the sorts of questions the algorithm thinks I might be interested in. The fact it sends them to my backup email account, the one I use for group chats and other stuff I don’t want littering my main inbox, suggests that I saw this one coming.

For a long time, the sort of questions Quora would fire back at me, along with the first few words of the top ranked answer, made a pretty good fist of being Elledge-specific clickbait. “Why is England considered a country while the UK or Britain, which it is part of, is also a country?” “Why can’t we dig deeper than 12.2 km into the Earth?” “Did Henry VIII regret executing Thomas More?” And once upon a time, I’m told, the platform was actually rather good: a place filled with scientists and historians, where you might actually find an actual expert to answer your question. 

In recent years, though, it’s gone noticeably downhill. The people most active on the platform today will show their bonafides not through their qualifications or professional status, but by listing the countries they’ve visited and the languages they speak. There’s something car-crash compelling about the tone of the site, too. In a manner familiar from food writers and also, I fear, this newsletter, answers to big questions will inevitably begin with half-remembered personal anecdotes, possibly dating back to the late 1970s. Quora taps into the urge many people feel, when exposed to an internet connection, to show off. It’s found a way to monetise mansplaining. 

These days, it sometimes struggles to hit even those intellectual heights, and in recent months I’ve noticed the questions I’m emailed growing increasingly, mindbogglingly weird. “Why is King Charles not called ‘King Charles the Ninth’?” (Well, why would he be?) “Why is Paris not the capital of France?” (Did I miss a memo?) “Why do Brits call their currency ‘quid’? What’s wrong with calling it ‘pounds’ like everyone else does?” (Oooh, that was so close to being an excellent question.)

The site doesn’t make it easy to identify who is asking its questions – to find out, you have to go from the email, to the website, to the page for the specific question, to the button marked details, which is probably more clicking than I’m going to do for anything that isn’t a video of two mismatched animals who’ve somehow become best friends – and the internet is full of people demonstrating quite outstanding levels of stupidity so I never thought much of any of this. 

Until, that is, I spotted an answer beginning with the phrase, “IDIOT QUORA PROMPT GENERATOR AGAIN!” Because it turns out, if you go to the bother of clicking from email, to website, to question, to details, a lot of these questions aren’t being asked by real people at all. They’re being asked by an AI. 

Not all of them, to be clear: the ones apparently aimed at winding up specific groups (Americans, Europeans, and so on) are almost certainly being asked by other specific groups (Europeans, Americans, etc). But where the questions become almost gleefully absurd (”What are people wearing in heaven?”; “Which K-pop idol has a crush on you?”; “What can we say to our cats?”) they almost certainly didn’t come from a human being at all.

The reason this is happening, the tech writer Kate Bevan told me, when I called her to ask, essentially, WTF, is almost certainly because “artificial intelligence” is a deeply misleading name.