The first chapter of Artificial Intelligence started with Humanisation, albeit in a broader aspect: one of the first Neural Networks was introduced by Kunihiko Fukushima from Tokyo in 1980.
He called it the “Neocognitron”. Fukushima proposed several supervised and unsupervised learning algorithms to train the parameters of a deep neocognitron such that it could learn internal representations of incoming data.
He was inspired by a paper published by the Nobel prize winning duo in medicine – David Hunter Hubel and Torsten Nils Wiesel which showed how the visual cortex processes light to generate vision. AI mimicks how we humans do basic things such as learn, see, talk and think.
And now, with Large Language models in AI invalidating Moravec’s paradox – where tasks hard for computers, such as learning languages etc. have been made simpler, thus one more barrier between humans and AI has been flattened.
The most critical element of Humanisation is plurality – and hence we need to look beyond Personal AI – what if Personal AI creates sub-objectives which are self-centric and disagree with plurality?
We need to ensure we create ‘Social AI’ as that is what makes it more Humanised. Let’s make AI ‘Augmented’ Intelligence for Humans – beyond that of Artificial Intelligence.
