AI for People and Business

Chapter 11. Scientific Innovation for AI Success

We now begin laying the foundational framework for developing an effective, vision-aligned AI strategy. This chapter is very important for executives and managers interested in using AI, particularly for strategic planning and appropriately setting expectations.

AI is a highly complex scientific field that is largely driven by exploration, experimentation, and unpredictable outcomes. As a result, you should think of AI as a form of R&D, including expectation setting and budgeting. AI is also a field that is very actively researched and undergoing continuous advancement, applied in increasing and varied ways and at a very fast pace.

I am regularly asked by business executives and managers what value AI can create for them (including ROI), how much time and cost AI solutions will take to build, what solution performance will be achieved, what AI technique or approach will work best, and what exact data is needed to ensure a certain level of performance.

Just as with many R&D initiatives, it is understood that people are usually unable to answer many of these questions upfront (that’s the entire purpose of R&D!) AI is no different, although I find that this is largely not yet well understood. The key point is that AI is a scientific field of discovery and not one of design and assembly. Let’s discuss why.

AI as Science

Let’s revisit the term “science” in the field of data science. Most of us remember the scientific method from school. ...