Garbage In, Garbage Out

I remember that back in the day, when computers still had novelty value and just to interact with them in the most basic of ways, you had to learn a programming language, there was a saying: 'Garbage in, garbage out'.  In other words, their output could only be as effective as the program you had inputted.  Mistakes in the code, or just downright poor coding would give you poor results.  The same, of course, is true of people - feed them a diet of false information and they'll harbour a false world view.  Propagandists know that well, which is why they make out sure that the media is full of their soundbites and stories which promote a slanted, out-of-context view of events, in order to try and give their world view hegemony in the public consciousness.  In recent times, the situation has become even worse, with those seeking to spread their dubious views now able to bypass the conventional media altogether, instead spewing out an endless stream of videos and clips on You Tube and TikTok pushing the crudest of hate-mongering - clearly directed at the young and impressionable.  But Hell, we know all this - I seem to have written about it endlessly - to little effect, obviously.  Now, of course, a new front has opened up with the advent of AI, which also is prone to spewing out misleading information, presented as facts and worse, it allows people to create realistic looking fake pictures and videos to try and discredit opponents and mislead voters.

Because, of course, that same maxim of 'Garbage in, garbage out' also applies to AI - no matter how much information a given system (illegally) scrapes from the web, if the sources it is using are bad, then it can only output bad answers.  Just look at Elon Musk's attempt at AI: Grok.  Informed, it seems, by what it reads on Twitter, which, under Musk, has become a cess pool of right-wing bigotry, racism, misogyny, transphobia and just about any other phobia you can think of, it spews forth 'answers' which reflect the general ill-informed  ignorance generally to be found there.  That, combined with apparent interference from Musk to ensure that his AI also reflects his own crackpot prejudices and you have a potentially very dangerous piece of technology.  But the 'Garbage in, garbage out' maxim also gives an opportunity to those of us who produce content to actively undermine AI systems and expose their limitations.  I'm well aware that my content here and on The Sleaze gets scraped by AI bots, which has motivated me to make out sure that I put out as much bizarre and utterly misleading stuff as I can.  I've made it my mission to try and ensure that the likes of ChatGPT start telling people that Boris Johnson was the 'Balham Buggerer', that London has been haunted by 'Flame Arsed Jack', a terrifying ghoul that lights its own farts or that Trump, clad only in soiled underpants, jumped out of a giant cake at Jeffrey Epstein's birthday party.  So, I urge everyone currently turning out content on 'little sites' like mine, I'd urge you to follow this example.  People have a tendency to perceive these AIs as somehow being 'neutral' or 'objective' arbiters of knowledge.  But they aren't: they are as prone to bias and prejudice as real humans.  But, if we can get them all to regurgitate our false information as if it were true, then maybe, just maybe, we can slow down their encroachment into everyday human activity and alter perceptions of them.  Either that, or we'll warp the minds of generations of the human race for decades to come.