AI has taken the world by storm and this time around the hype created around is so loud that everyone seems to be overwhelmed with it. People are just too scared of the FOMO effect! What if I don’t ride this wave? I’ll be left behind. Unfortunately, there are significant and important aspects of AI that most people are either not aware of or have chosen to ignore them for various reasons.
This is not nay-sayer’s argument, rather I’d like to draw your attention to the pitfalls of this technology space and where is it failing to generate the purported benefits.
AI is Not New…
AI has been around for a while now. In fact, it is probably older than most of the technologists who are now playing in this space. History of AI shows that this field of study goes back to the 1950s! That’s 70+ years from now. The first thing that I’d like to draw your attention to is the lack of ROI (and not only in terms of money, but in general) that AI has been able to generate over the last 7 decades. It has promised but not really delivered. 7 decades is a long time for any technology to mature and deliver on the promise. And one must ask the question why? What is missing? Where are we going wrong?
Ahead of its time back then…
When AI first burst onto the scene, it was ahead of its time. There wasn’t enough computing power and resources available to make it practical. Yes it kept brewing in the annals of the academia and research institutes, but that’s where it largely remained. Out of the common man’s reach. This has gradually changed, and now we are seeing the “democratization of AI”, such that it seems to be available to the common man at large. Thus while it didn’t provide the benefits back then as expected, why is it still failing to do so?
AI is a probabilistic system…
We have either forgotten or are overlooking a key characteristic of AI. By its very definition AI is a “probabilistic” system. It will never be 100%. It will only be correct to a degree. It is foolhardy to assume that AI will be perfect. And one must analyse, why has this expectation come to be? The simplest reason is that computer science by its very nature is deterministic. 1 or 0. True or False. There is no ambiguity in between. Either a program works, or it doesn’t. At the end of the day, even an AI implementation is an algorithm (a software program) that is working on data model to generate its outcome. And naturally enough, every one expects it to be “right”. To be correct all the time. What would we do with an incorrect or partial response? And this is the dichotomy we are grappling with.
“There is NO intelligence in Artificial Intelligence”
This might seem like I am preaching, but remember, the outcomes of an AI system is completely dependent on the data. If the data is wrong, or if the data is skewed towards a particular outcome, that is what you will get. If data says, it is a good idea to jump off the cliff, the program will jump off the cliff and gladly at that, if I might add.
In a nutshell…
There is value in this technology, and we are all hopeful to see that materialize, but there is no denying the fact that we are burning truck loads of money will a return on that investment that is not justified. I urge leaders to be pragmatic and deliberate about what would really work and won’t. Where is the true bang for the buck? It will require R&D and perseverance but don’t be like a deer on a highway blinded by the headlights of an oncoming car… accident will be inevitable.