Last week, an expert panel under the auspices of the National Association of Corporate Directors–New England, presented on what a Board should know about use of AI within their own corporation.  To this observer, perhaps by reason of awaiting specific guidance to be passed on to Boards, the session was fascinating but somewhat frustrating.

Consultants and investors agreed that at this point most entities were at a very early stage, experimenting and not going “all-in” on any function wholly dependent on AI beyond very basic data analysis (for which the less sophisticated AI was wholly adequate)..  Many experiments were often “fails” on a learning curve.  It was estimated that no more than 5% of companies were deep into AI implementation for a variety of reasons, including fear of the biases or hallucinations  that might arise from AI relying on large language models, and fear that one’s own AI might find its way into the hands of competitors.

It also seems that the key at this stage is for a Board to identify “use cases” in which to experiment with an AI solution.  Little guidance was given as to how a Board is best able to identify its own optimal use cases.

While no one on the panel stated this, it seemed clear to me that a board member without specific application expertise would be at a loss to have an inkling as to initial direction.  This observation leads to the use of expert consultants, or larger enterprises hiring a group of AI experts; it was not clear where a group of AI experts might be found, given the lack of industry experience noted above indicating a good direction.  Indeed, the panel itself seemed to possess substantial insight into possible approaches, but the panel was culled from elite experts not likely to be found on the average (albeit highly competent) Board.

It should be noted that in the past couple of weeks Economist magazine also noted, in a major cover story, that actual business use of AI in almost all significant use  cases was quite unusual; there seemed in the coverage some ROI hesitancy in the level of investment capital today going into the AI side of the economy.

While I have generally not been a fan of adding a Board member for each vertical function, favoring general board experience or domain experience in the enterprise business space, I confess to having concluded that AI at this point may well be a spot for one AI-expert director given the risk and the size of the spend involved.   (Until now, I was of the view that skill-specialized directors beyond the specific business involved were highly desirable only for financial control/accounting functions in certain kinds of enterprises.)

One thing was quite clear: all on the NACD panel were quite sure that at some point AI would be integral in a major way in all business, with continuing  huge investments coming into the AI space.  In AI, the best is yet to come– with the caveat the worst to come may be the next step.