Everything that has become a norm in our lives went through some hype cycle. So in essence people may overblow its usefulness and think these things are going to change the world but then it doesn’t change the world overnight so things comes crashing down for a while before it goes on to slowly change the world. The internet is probably the best example. In the 90s people were sure that the internet was going to change everything, and so it went through a bit of a hype. So did computers in the 80s. But after the hype, things crashed, and then life went on except it got changed bit by bit, steadily and surely.
Generative AI is itself going through a bit of that; we are all sure things are going to turn out great. Of course, some doomsayers will be warning the world of the problems and calamity it would bring – just as Socrates thought writing was a poor form of communication and would also bring about the decline in memory of men. I think it is probably necessary to create more safeguards for AI and allow the governance to evolve with its development.
I think Gen AI will be helping to augment the capabilities of human workers for a really long time before they come to ‘replace’ workers so to speak. Yes of course you could use some kind of AI technology to help you even have a conversation at the call center, but it’s not going to be able to handle 100% of the queries, you will eventually still have a human in place. Consultants for example, who might have been spending time copy-pasting or doing data entry type of work might lose their jobs but then there will still be someone senior who needs to intervene.
The real economic challenge for us is how are we going to let Gen AI do the so-called low-level jobs while maintaining a pathway for us to train more junior workers into capable senior workers. Sure there is the grunt work that has to be done but traditionally, the juniors learn the ropes by doing those work. If they are going to be performed by Gen AI, then how on earth are they going to be able to get the chance to learn?
There is still substantial job opportunities which are slightly underpaid but cannot be replaced by Gen AI. These work are underpaid either because of systematic biases in the economic systems or as a function of labour market rigidities. They include the care-giving, pastoral guidance type of roles, as well as all of the cases where it is important to have a human example who can model moral character and other crucial human attributes. No kid is going to see the politeness of a Gen AI figure or speech bot and say he or she wants to be courteous because they are a role model for the kid.
To me, those problems will need to be gradually resolved before we would allow AI to play a bigger part in the lives of people. Part of the way some of these problems are resolved is actually through mutual cancellation with the demographic transition challenge. Economies that are mature and have severely ageing population will need to rely on AI for many things. Improving labour mobility globally should slow down the need for that but it is inevitable for these markets who have the resources to play the early adopters’ role.