Deploying Artificial Intelligent Projects — Struck by Definition
Artificial Intelligence (AI) is gaining mainstream acceptance rapidly, in stark contrast to just a decade ago, when getting funding for AI projects was a big challenge. The extend of AI’s prevalence in our daily lives is summarized in the poster below, from the moment we wake till we sleep. I know this may come as a surprise to many. Whether we like it or not, the convenience afforded by AI is a clear indication it is here to stay. This is just the beginning, with the application of AI limited only by our own ingenuity.
So, what’s up?
Despite the hype, the recognition of its importance, and the so many success stories you read in media, AI adoption at the corporate level is miserable even today. Depending on which report you read, successful deployment rate of AI projects has been reported to be between 10% to 50%. There are various reasons for this.
Working from home now means that I have got the time now to pen down my thoughts and musing on the state of AI in the industry. This is the first part of the many that is to come as I share my experience deploying AI solutions.
Struck by Definition
It might sound trivial but there are simply too many differing opinions on AI which is disruptive. Adding to the confusion is the deluge of buzzwords such as machine learning, GANs, deep learning, and so on. There is a very popular (but technically misleading) tweet that reflects the situation:
“If it is written in Python, it’s probably machine learning. If it is written in PowerPoint, it’s probably AI”.
I will start off with Hollywood’s definition of AI (since it probably has the strongest influence on many people out there). Here it is…
And then we have some of Academia’s definition here…
The definition by Keene is my personal favorite because it reflects reality. This phenomenon is also known as the AI effect. Humans, due to our need to make ourselves the centre of universe, have continuously shifted the AI goalpost further and further. Meaning what was accepted as AI last year may not be considered AI now.
Why is it so important to make this point?
Because I have been in weird situations where workable solutions are jettisoned or failed to be deployed because it did not meet the customer’s expectation of AI. Because I have been in so many discussions centered around the precise wordings, nuances and I have not seen any impact of the “Aha” AI definition.
Definitions do not matter when it comes to improving business efficiency or disrupting your business. Focus on the outcome that AI can bring to the table rather than getting too engrossed in getting that sparkling definition. Please do not allow such trivialities get in the way of your digitalization and AI deployment exercises.
And this will bring me to my subsequent posts which focuses more on the Human Factors side of things.
