Overuse of AI - Risks and Blind-spots

Analysis paralysis. AI excels at pulling together and interpreting large bodies of data. It is good at identifying anomalies, finding patterns, and making predictions. Because it’s so easy to use, some executives are seduced into using it as a substitute for their own critical judgment, or they fall into the trap of over-analyzing a situation, and not actually making a decision.

Loss of contextual understanding. AI is still “narrow” in scope, meaning that on a specific assignment or task it can be every bit as smart as a human, but without any ability to tackle slightly different tasks. The increasing use of AI in the business world is therefore helping executives to optimize their answers to well-defined questions, while potentially reducing their ability to see the bigger picture. Many investment decisions, for example, are based on Net Present Value analyses, which means that if a potentially important factor cannot be quantified, it doesn’t get included.

Lack of differentiation. The race to develop new AI technologies is fierce, which means that advances are quickly copied. Of course, some of the firms developing AI technologies generate a temporary advantage, but their business model is typically to embed that technology in as many of their client firms as possible (e.g. IBM’s Watson business unit). This leads to a predictable “arms race” where competitors invest in very similar technologies, to avoid falling behind, but they end up even less differentiated than before. The fund management industry, for example, increasingly uses “robo-advisors” to make investment choices for their clients, but competitors all have similar algorithms for making their trading decisions, they will inevitably end up with very similar investment returns.

While AI can help achieve unexpectedly good outcomes, it also has a flip side. Giving an extraordinary exalted status to AI should not undermine the critical reasoning of inimitable brain