The Balancing Act

Perspectives surrounding our own often get lost and forgotten, leaving us narrowly focused on only one side of the multi-faceted world. In the past few weeks, exploring advancements and experiments conducted with Neural Networks has been exciting, filled with lists of questions and mouthfuls of new concepts. However, it has been far too easy to ignore the real world and the reality behind Artificial Intelligence in a career setting.

This week, however, my eyes were opened to a few of the significant perspectives impacting the field of technology in general, but specifically applied to Machine Learning. I recently got the opportunity of listening to an interview with Sri Annaswamy, the Founder and Director of Swamy & Associates, by Verint where Artificial Intelligence and its prevalence and future regarding customer service and call centers were being discussed. What truly stood out for me was the fine balance between the Business and Technological perspectives and how different they may be, yet how interdependent they are. Technology requires funds and support, while Business cannot efficiently run with obsolete systems and methods.

Oftentimes, I forget to consider the practical industry-side while trying to study or learn more about this field because I mistakenly do not realize its importance. One of the examples that was discussed in the podcast included pilot programs or experiments conducted in small-scale environments to evaluate the practicality and feasibility of a real model, which could be scaled up in the future. Unfortunately, many of these pilot programs often never reach a wide-spread level due to a lack of support. The issue, according to Mr. Annaswamy, is the lag between Business expectations of AI, consumer expectations of AI, and the reality of AI. So, the misconceptions with how simple and easy AI is supposed to make profit-making, leads to low support and resources that are required in the initial stages of building and training effective models. The world's need for quick, low-cost results often overlook the costs associated with not being able to scale-up models, or having to start all over again. The effort needed for long term activities may be higher in short term, but could be significantly more cost effective compared to dealing with models that use unrepresentative data, or mislabeled data, or improper training, etc. For this reason, technological advancement could potentially be stunted.

In addition to the evident gap between ideals and reality, what struck me as bizarre is with all the attention that the field receives, a large portion of factors related to AI fluctuate as a result of public opinion. Media tends to boost expectations for AI and optimistically paints a picture that our present technology is not in a position to achieve. Therefore, it seems that AI is within the clutches of media and people's speculations on the industry and not because of true, scientific advancements.

I realized the importance of not only including the Business and Public face of the industry, but also how crucial of a role communication plays in blending these different perspectives together to effectively arrive at optimal solutions. It is inefficient to become experts of all areas, which would defeat the purpose of specialization, yet it is important to step away from misconceptions or misrepresentative generalizations in order to make progress. We should strive to expand our understanding and clear communication of different sides - the handshake of perspectives.