Elon Musk and the age of the troll

There has been a cheering new development in the struggle against scam phone callers. AI can now be used to automate the satisfying but tricky business of ‘scambaiting’. I give you Daisy, the ‘AI granny’ – whose only purpose in life is to keep phone fraudsters on the line for hours that they would otherwise spend predating on real human victims. Scammers, as we know, play on human psychological weaknesses – the panic we feel when we’re told our accounts have been compromised, our deference to authority, our confusion about how technology works. Now, flip-flop: the AI plays on scammers’ psychological vulnerabilities – primarily, the idea that an old lady with a quavering voice will be an easy mark.

The phone company O2 has devised an AI programme that supplies a convincing telephonic facsimile of what you might call a Batty Old Doris. The scammer, convinced he’s hooked a dear old lady who will easily be relieved of her life savings, attempts to ferret out her bank details, date of birth, passwords and what have you. And Daisy, indefatigably, burbles on about scones, amateur dramatics or her love of knitting, and offers confused versions of recipes for lemon meringue pie for as long as the poor scammer can endure it. The clock ticks on and, hours of time having been spent down the fraud mines with nothing to show for it, it’s pleasing to think the scammer can expect a roasting from his supervisor when the daily figures are in.

Why is this such a very pleasing development? In part, it’s because we all know or can imagine some version of the Batty Old Doris on whom Daisy is modelled. We can feel an intense dislike for the pitiless cruelty of the scammers who, when connected to a flesh-and-blood pensioner, will intimidate and confuse her and, most likely, steal money she can very ill afford to have stolen. It’s a vicious and heartless criminal enterprise, and with no real prospect of the scammers being meaningfully brought to justice by any police force the next best thing is to waste their time and energy, frustrate their efforts and ruin their days. It’s a fitting revenge.

But there’s also a special sort of delight in something else. This is not criminal justice as anti-fraud measure; it’s an automated version of what’s sometimes called trolling: that addictive internet recreation of winding up strangers online just to revel in their reactions. Nobody, I dare say, is completely immune to its appeal. As it has moved from the margins of the internet in places like 4Chan and Reddit to the mainstream, there’s a case that trolling is becoming the dominant form of public interaction of our era.

The most interesting and distinctive thing about Elon Musk and the age that has sent him to prominence is not his riches or his ambition or his political heft. It’s the fact that he conducts his business empire (and intervened in an electoral campaign) like the most basic internet shitposter. He’s all about the lols. Catapulting a car into outer space; offering $420 a share for Twitter (420 being a slang reference to cannabis); joking that a ‘prophecy’ would be fulfilled if Trump polled 69.420 per cent; and finally naming a government department, DOGE, after a meme cryptocurrency.

At the centre of the US government will shortly be somebody whose mode of being in the world isn’t in any important way different from that of the YouTuber Mr Beast. The instinct is for the eye-catching stunt, the callow meme, the silly joke, the wind-up. And what got him there was, in part – especially in the very online communities where the noisiest part of the election was played out – a programme of action that was less about what you believed in than about what would cause maximum horror to your opponents.

The online right lives to ‘own the libs’ and drink ‘liberal tears’: these are not pleasing by-products of the programme so much as the whole point of it. The real players there aren’t Chicago School economists or hawkish foreign policy academics: they are an army of Pepe the Frogs and Bitcoin fundamentalists with NFTs as their Twitter avatars. And what, come to that, was Russian state TV running naked pictures of Melania Trump but trolling as an instrument of foreign policy?

I don’t point to all this just to disapprove of it. I’m old-fashioned enough to think that Musk’s unseriousness might carry risks, as might – for instance – an incoming president picking his cabinet on the apparent basis that the best choice is always the one that will cause the New York Times to have a nervous breakdown quickest. That ship has very much sailed, and unless we decide as a species to turn the internet off for good we’re just going to have to get used to trolling being at the centre of our politics and public life, just as we’re going to have to get used to AI being the unstoppable pollutant of discourse that it is. We’re going to have to find ways to negotiate those two things. And that is where Daisy comes in.

It was the satirist Charlie Brooker, I seem to remember, who described some years ago why he gave up playing Scrabble online. Irritated at being beaten so often, he had started keeping another tab open while he played, with a program which would tell him the optimal play in any given situation. He’d been doing this for a while when it occurred to him that most of the people he was playing with were almost certainly doing the same thing. Essentially, two rudimentary computer programmes were playing Scrabble with each-other, and Brooker and his friends were spending a significant portion of their only lives on earth gormlessly playing the computer’s moves for them.

This depressing realisation also, I fancy, allows us a chink of hope – at least when it comes to technologies like Daisy. After all, it can only be a matter of time before the scam factories are using AI themselves – as being cheaper to hire even than third-world call-centre operatives, and better at tricking people out of their bank details. Eventually, we can imagine, millions and millions of scam phone calls will be going on all over the world every day – and a figure approaching 100 per cent of them will be calls between one AI trying to get bank details from another AI who has no bank details to give.

Indeed, it’s perhaps not too much to hope that within a year or two we will all be able to outsource our shitposting and ragebait on social media to AI too. I might, for instance, ask ChatGPT to set me up a Twitter account which would tirelessly post outraged tweets of the sort that a gloomy milquetoast liberal might post, and it could get on with farming engagement by starting shouting matches with the bots impersonating Laurence Fox or Allison Pearson. Eventually, all the major social media sites would be entirely populated by bots trying to wind each-other up and trick each other into clicking on their OnlyFans links. 

And at last we humans can breathe a sigh of relief, leave them to it, and settle in a comfortable chair with a slice of lemon meringue pie.

Watch more like this on SpectatorTV: