Points to remember when embarking on an organization-wide turn to AI solutions

In our organization, we are currently in the phase of building up team, skills to automate and implement AI based solutions. So, we are very early in this AI journey.

Right now, we are also working on identifying some of the problems that we face in our business. For example, when we get 8 customer segments, but only 2 of them bring in a lot of revenue. Rest all of them perform poorly. We would like to find out why through data analytics/identify factors that is causing this issue.

While all this seems doable, I would like to seek your suggestions on how can we make the business users/leaders clear on what AI can and cannot do. Because, I feel it is very much possible for business team to be carried away by the hype around AI/ML etc. So, as a data person, I think it is my responsibility to clarify what can and cannot be done using AI. And why can't we rely on AI results 100 percent. Why should there always be a caution in trusting AI output

Any books,papers, case-studies or articles etc which has this information/points to consider when embarking on organization wide AI-initiative can help me

One such article is here

Topic ethical-ai deep-learning neural-network machine-learning

Category Data Science

1
answered at

Some important points I can bring up are:

  1. AI/ML learn (stable) patterns from what they are given. But, if forced, they will find (irrelevant) patterns even in noise.

  2. Cannot learn what they did not see. So, generalisation is actually possible only for variations (allowed by the underlying architecture) of what were already seen.

  3. AI/ML may discriminate and not be fair, where fairness is required (this can happen for various reasons).

  4. It is not always interpretable, so one cannot know why one gets this or that result. More importantly one cannot verify (in a straightforward manner at least, before the fact) if any of the previous issues happens.

References:

  1. Overfitting
  2. Generalization error
  3. Concept drift
  4. Ethics, transparency and accountability of AI
  5. Explainable Deep Learning: A Field Guide for the Uninitiated

About

Geeks Mental is a community that publishes articles and tutorials about Web, Android, Data Science, new techniques and Linux security.