AI can be your worst enemy

The recent viral sensation of just about everyone using AI to create caricatures of themselves that are often eerily accurate, as well as telling the rest of us how they see themselves, has been fascinating – and fun. If you haven’t yet, you can go onto an AI platform of your choice, upload photographs of you and then prompt the app. If you haven’t seen it yet, do yourself a favour and find the one of President Cyril Ramaphosa in the form of a drinking game that was doing the rounds in the lead-up to the Sona last week.

Players had to take a shot of vodka for every time the president used the word “shocked” or “appalled”, a sip of beer for “we take this matter seriously” and a glass of water for “working together’. It’s good fun, using tech for satire – but what else is it? AI is premised on machine learning.

It needs info to crunch. Rene Descartes famously said “I think, therefore I am”, but we were are deep into the dopamine tick era of “I’m liked, therefore I am”. There’s nothing wrong with that at all, but the quiz takers – and now caricature fans – are not too far divorced from the WhatsApp group that moans about the blindness of “main stream media”, or MSM for not reporting on the stories, while sharing the mini-URL of the actual report – or those who trust the sheep dip, ivermectin to cure every ailment from Covid to cancer and even autism.

[paywall]

AI is an amazing invention and our lives are immensely enriched because of it – but there are also inherent dangers in giving away information to help it understand us and our likes and dislikes, even better than we do ourselves. It’s one of the greatest ironies that the prevailing complaint of the keyboard worrier class is how the tech elites have too much power over the ordinary individual, little realising that every time someone takes part in a seemingly innocuous quiz, or creates a cutesy caricature of themselves, they are voluntarily surrendering the information needed to be weaponised against themselves.

[/paywall]