All my life I have been fascinated by car ferries. For decades this has been a seriously informed interest.

I have yards of books about them. Can bore for Britain about turntables, bow-loaders, side-loaders and hoist-loaders.

Lecture on Voith-Schneider thrusters till you look desperately round for a weapon, and have never met a slipway I didn’t like.

So why, these days, and when I visit Facebook, am I immediately regaled by footage of ferries going horribly wrong? Cars drive off and straight into the drink. 

Pitiless Atlantic billows rage through vehicle-decks; rearing linkspans flick trucks into the air like medieval catapults.

On closer scrutiny, though, these glimpses into a world where, seemingly, bad ferries go when they die do not quite add up.

The shouted commands of pier-hands (usually, in polite Australian) make no sense. Folk do not usually drive straight off a boat and into the briny like witless lemmings.

This, of course, is a world that is not so: twisted fever-dreams from Artificial Intelligence.

The creepy thing is not the knowledge that there are bots out there capable of creating, effectively, snuff-videos starring that venerable CalMac veteran the Isle of Arran – which began her career when I was still in school and is so old her lifeboats have oars, though reports of her participation in the Battle of Lepanto are but foul calumny.

The point is that out there, somewhere, AI has specifically locked onto me, in this unfolding nightmare of drowning Audis and motor-munching bow-visors, because it knows I have a thing for ferries.

As Nietzsche darkly put it, if you gaze long enough into the abyss, the abyss will gaze back into you.

For, every time you visit the web, send an email, cruise the information superhighway or hail the cosmos via the-artist-formerly-known-as-Twitter, forces beyond our conceiving are stuffing an ever-fatter dossier on you, your life, your interests – and your darkest secrets.

It sounds incredible, but Facebook knows you are falling out of love with your partner before you do. More prosaically, we have all had the experience of Googling, for instance, ‘What is the best omelette pan?’ – and being stalked thereafter online, and for weeks on end, with adverts for omelette pans.

Facebook is particularly good at this sort of thing because, though relatively late to the social media party, it has been so brilliantly run for so long it has seen off the forerunners: Friends Reunited, MySpace and Bebo are as Nineveh and Tyre.

In public life, now, whether you want to or not, there is immense professional pressure on authors, journalists and politicians to have an ‘online presence’.

It can be sapping. In his last, 2015 election campaign, such was the online abuse heaped on the late Charles Kennedy – grieving and unwell – by the worst CyberNats that it was the full-time job of two people to run all over his socials tearing posts and comments down.

He lost his seat: a month later, he was dead.

In the folklore of my old school, a venerated headmaster was in 1952 beset by quite a few anxious parents. They were eager to acquire a television set, but should they wait until Roderick and Fiona had passed their Highers?

As television troubled many then, now many thoughtful people worry, aloud, about Artificial Intelligence.

Indeed, Pope Leo himself – at the Vatican’s first Mass, last year, for social media influencers – dropped words of warning. ‘Nothing that comes from man and his creativity should be used to undermine the dignity of others,’ he declared.

He also warned of a ‘naive and unquestioning reliance on artificial intelligence as an omniscient “friend,” a source of all knowledge, an archive of every memory, an “oracle” of all advice…’

The internet is not quite the free and unpoliced space we think it is. During Covid, Facebook added clunking riders to any posts or comments that made it nervous about social distancing or vaccination.

Serious allegations about Joe Biden’s son were oddly muffled, online, in the runup to the 2020 Presidential election – and, to his understandable fury, Alex Salmond was in 2019 ‘unpersoned’ on the official SNP website, all mention of him wiped from the party’s online potted-history.

But AI is another order of menace still. It has become a nuisance on the internet, Google and the rest immediately uploading AI results upfront when you search for anything. And you will almost always find fuller and more reliable information on Wikipedia.

Too often, in fact, AI’s leaden pronouncements are little more than gobbledygook. Yet you turn to YouTube, spot brilliantly paced, deftly acted and well-lit little films – and then spot it’s all AI.

AI ‘journalism’ is now such a pest that most publications now deploy software to sieve it out from freelance offerings.

Indeed, there was a brilliant article about this in The Critic – ‘Stop sending us AI articles: we can spot the slop,’ last June. But when I tried to find it online just now, AI Overview corrected ‘slop’ to ‘slope’.

‘They argue that AI-generated content (or “slop”) is recognizable, formulaic, and inferior to human-written work,’ AI Overview pronounced stiffly, ‘often criticizing the overuse of AI in publishing and media.’ AI does not do irony.

There was an eerie Doctor Who forerunner of all this 50 years ago. In the 1976 storyline, The Deadly Assassin, the Doctor – of his own volition – is put to sleep and plugged into the Matrix, a vast ‘database’ of collective memory and pretty well what the Internet is now.

He finds himself in a nightmarish reality of cackling clowns, stumbling Great War soldiers, ponies in gas-masks, railway-points that grab him by the ankle as a train shoogles towards him, and, ever closer to him on the trail, an implacable killer…

It ends on a freezeframe cliffhanger so terrifying Mary Whitehouse kicked up a storm and the BBC subsequently edited the mastertape.

But that hallucinatory realm is, half a century on, the world we now live in.