Sunday, March 04, 2018

The humbling experience


I recently attended a meeting of a small group of technology and domain experts. The meeting was to focus on a problem statement and dissect it under different technology and domain lenses. Each person was from a different field and bought a unique expertise to the table. It was evident they had spent years in their focus area assimilating and growing their knowledge. Very quickly I had realized ‘expert’ was too modest a noun to describe them. The meeting which had started as an amazing experience had soon turned into a humbling experience for me.

I have been thinking about that experience for the last few days and seem to have found a correlation between that meeting and Expert systems/Artificial intelligence-Machine learning (AL-ML) system. That meeting to me resembled an expert system building pattern where experts pour their knowledge and experience to solve a complex problem.

As I thought more, my mind started drawing random parallels between that meeting and expert systems and generally wandered into the direction of the expert systems. Every expert system (or an AI-ML system) when introduced, first sparks a sense of bewilderment. It is approached with a set of presumptions and questions, bordering disdain. As the expert system proves and grows with promise, the disdain soon turns into an admiration for the expert system. The user starts appreciating the system and its potential. Often the admiration might even give way to distrust. A distrust might arise from a question of job security or individual well being or even from a question of ultimate survival. Take the example of Self-driving cars; they were initially struck down as a work of fiction. But as the self-driving promise held ground, it turned into an admiration. Distrust too found way in the hearts of the car enthusiasts who hated the very thought of sitting in the back seat and not at the wheel.  Isn’t it a similar pattern for most AI-ML systems, maybe with only slight variations? But one stand out feature among all these systems is the sense of humbleness that they evolve. Once you accept and understand the power of expert systems, they undeniably lead to a sense of humbleness. The user is humbled at the sheer talent of these systems.

As my mind searched for parallels, some things started falling out - the non-parallels. The non-parallels in terms of human emotions, or spontaneous thoughts, or even simple non-verbal expressions. Are there parallels to these? Even two humans can rarely experience the same levels of emotion. Any level of admiration, distrust or even humbleness that the AI-ML systems evoke, cannot replace these simple non-parallels. It is these subtle non-parallels that defines us as humans, and separates us humans from the AI-ML systems.  Maybe at some point in the future when the AI-ML systems have significantly evolved, they too might hit a point of humbleness. The humbleness that might arise in them from the fact that non-parallels exit between them and the humans.

~Narendra V Joshi

No comments: