When I told my dad I was going to be an anthropology major, his doctor patted him on the back and told him, “Oh Henry, Susie just bought herself a one way ticket to the poor house.”

I reminded him of that when my sister and I bought my parents their retirement house.

A recent Wall Street Journal article by Callum Borchers made a surprising recommendation for the AI era: study anthropology.

As an anthropologist and innovation storyteller, I was delighted, but the reason matters more than the recommendation itself.

The article quotes investor Alec Litowitz, who studied mathematics and anthropology at MIT, describing one of anthropology’s greatest gifts as metacognition: the ability to step back and observe a system while recognizing that you are also part of that system.

That distinction is easy to miss. It is also one of the most important capabilities an innovation leader can develop.

When we study customers, employees or markets, we tend to imagine ourselves as objective observers. Anthropologists are trained to question that assumption. 

What am I bringing into this observation? Which assumptions are shaping what I notice? What feels “normal” simply because I have been inside this system for years?

Now put AI into the picture.

A Fortune 500 company can feed an AI system millions of customer interactions and still misunderstand its customers. Data reflects the system that produced it. Historical behavior contains historical assumptions. Organizational data is shaped by what the company chose to measure, what customers were asked and what employees were rewarded for.

AI can find the pattern. It cannot automatically tell you whether you are looking at the right pattern.

That is where anthropology becomes strategically useful.

This is why I think the future of innovation leadership requires something more sophisticated than AI literacy.

It requires AI plus anthropology. And I never thought I’d see that in the Wall Street Journal.

AI gives us extraordinary powers of observation. Anthropology reminds us that observation is never neutral.

The biggest risk isn’t that AI gets the story wrong.

It is that AI gets the story right because we asked the wrong question.