if you are a fan of Gartner, and who among us is not – right?, you know about the hype cycle.
image above, copyright Gartner, Inc.
funny, well – i think it is, was that when i used to present as a G analyst i used to call it the marriage cycle… you meet someone (trigger), you date them for a while and you think they are the ones you need to spend your life with (peak of inflated expectations), you get married and — poof! reality kicks in and you find yourself in the through of disillusionment, only to little by little emerge into a working marriage later in the plateau of productivity (have kids, raise them, put them through college, etc.). the line that killed it was when i said after marriage you sink into the through of disillusionment — guess you had to be there… but i digress.
the theory holds for technology adoption, and we have recently been hyping AI as much as we possibly can (for the 8th time since its inception, but that’s a different blog post). and i have been looking for signs that we are getting to the top – because that is the best place to be (IMO). you see, if you can no longer hype a technology then you can start looking for applications (real-life use cases) and start learning how to leverage it better.
there’s a ton of value that AI can provide to enterprises – but we are nowhere near there yet. this is why i want to stop hyping it, start looking for useful applications for the technology that grow the pie and move things forward.
and – usually – you find the source in the most unusual places. social media, for example, found its peak around the time that great recession was beginning to pass, mostly because people now could focus on things they needed to do, and had money to do them with. for AI, it is more complicated for three reasons:
- this was the eight generation, so there is lots of fuel (read, history) to power hype
- technology evolution, what brought AI to the forefront again, was truly a game changer this time around making it more powerful and more interesting
- because it is such a established discipline, and the tech was really powering progress, it was hard to distinguish hype from reality (GAN, i am looking at you…)
alas, i found an article today while perusing things that got me thinking we are there.
AI, beyond the technology used to implement theories, is powered by cognitive sciences (the science behind learning – as in, you cannot teach that which you don’t know or understand, and as computers need to learn – you need to know how to teach them to learn. this was very clear in the early generations before powerful computers). virtually all methods used in AI are learning methods we transformed to algorithms (read, programs or code, if it makes it easier).
the article i found talks about the answer to teaching computers better lies within kids – which is not only true, and a basic tenet of AI, but also the first article in a long time (non-academic) that recognizes that AI has not lived up to the hype and it may be time to go back to basics.
now, with peak of inflated expectations in our rear view mirror we can finally accelerate towards the trough of disillusionment (sorry, not curing cancer this time either – but we are making significant strides in ways that were not possible before…. one day… one day) and find appropriate ways to use AI in the enterprise.
until then, read about cognitive sciences holding the key to AI and marvel on the warm embrace of reality.
article can be found here, for more details.