No to Citizen Data Scientists, YES to AI Business Professionals!
by; Mark Stadtmueller, VP, Product Strategy
AI Summit in New York last December was a great event. In one of the sessions it was recommended to read “The Book of Why” by Judea Pearl and Dana Mackenzie. I finished the book last night. It was fantastic. I admit, some of it went over my head, but the style and storytelling made for a great page turner.
But, when I closed the book, I thought...was that all of us in the Enterprise AI industry? Are we suppose to stop talking about Citizen Data Scientists? Not only is it a bad term, but it is potentially very harmful. I think it similar to suggesting that we should have Citizen F-35 Fighter Pilots.
I understand the concept whereby Enterprise AI represents a singularly unique way for businesses to grow and differentiate themselves. The current (completely unworkable) model is for businesses to hire (or hope to hire) brilliant data scientists and then wait for some magic to happen. But, while waiting for the hires/magic, innovative people take innovative courses like Jeremy Howard’s fast.ai courses or Andrew Ng’s Coursera course sequence. In those courses, magic really does happen. The insight from these courses and the power of open tools that are employed in AI, people from every walk of life learn to do some amazing things. Regardless of background, people willing to code, think, and work hard can become AI Practitioners.
AI Practitioners with some business know how, can quickly start adding value in business. Individuals with a lot of business know how and keen insight into how AI and Data Science work, can definitely become an AI Business Professional. Somewhere along the way, it was decided to call them “Citizen Data Scientists”. It probably comes from Information and Communications Technology people who started talking about “Shadow IT” (i.e. IT done outside the IT department). But, here’s the thing, it is not the same thing. Someone who learns AI and knows business in not anything like a Data Scientist. I challenge anyone to read Pearl’s Book of Why and not conclude that the idea of a Citizen Data Scientist is not only wrong but potentially harmful.
Data Science is Science and while a Data Scientist can certainly learn to leverage AI, trying to make Data Science assertions without Data Science training and education, causal hypothesis and understanding, and skills in the scientific method can lead to a lot worse results than biased or bad data can.
As long as we understand that, the flip side is, a lot of businesses think everything to do with AI is Data Science. Businesses try to use the hire data scientists/hope for magic method, when they should be focusing on how to gain value from their data. However, now, knowledgeable business professionals, along with the tools necessary to easily and securely access and explore data; munge and yes try out AI models; and then easily pass on to serving production systems, are able to bring value from the data. The models may have even been built by data scientists or the AI Business Professionals may have consulted with data scientists. And even yes, an AI Business Professional may even be a data scientist. But, nowhere should there be a Citizen Data Scientist. Data Science is performed by Data Scientists and F-35s are flown by pilots.
Citizen Data Scientist: No. AI Business Professional: Yes.