The Blind Men and the Blind Elephant

In the old story about the blind men and the elephant, each man stood near a different part of the elephant and described what they thought the elephant was. The one who stood near its tail pronounced the elephant to be a rope. The one near a leg was sure that it was a tree, etc. The various perspectives on artificial intelligence are often the same, with just as much under-informed confidence that the blind men had about the nature of the elephant.

Image from ChatGPT

There are people who use generative AIs to do calculations, something for which they were never intended, and they write the tools off as totally useless. Others use them as search engines, another function that they do poorly. To them, all AI tools are undependable. They ignore the fact that the AI "elephant" is also "blind" and is not able to differentiate between fact and fiction. Still others briefly try the tools without spending the necessary time to learn their weird ways and they walk away without gaining the skills needed to use the tools properly.

The breathless marketing of the various retail AI companies hasn't helped. They have positioned their products as magic solutions to everything from healthcare to halitosis. They also tacitly allow people to believe that AIs are general intelligence tools at the level of movie androids. No wonder new users are disappointed

But in many classrooms and offices, the adventurous are pushing past the popular misconceptions and creating new use cases that make their daily jobs easier. In many cases, these adventurers are working under cover in organizations that discourage or outright forbid the use of AI tools. They are the Secret Cyborgs from Ethan Mollick's book, Co-Intelligence. These are people who are inventing the age of AI under the radar of their managers and often without the knowledge of their co-workers. They have removed their blindfolds and adopted and adapted AI tools in ways that their inventors could not have imagined.

In huge companies like Amazon, (I just finished a summer fellowship there.) AI is being broadly encouraged through company-specific tools that are being customized for users through direct requests to the developers. Teams are reporting increases in productivity even very early in the tools' adoption.

That brings up another aspect of the elephant. As a productivity rocket, it is also a threat to employment. Productivity increases have been essentially the only reason for corporate AI adoption from the beginning. Nobody should be surprised at demanded higher workloads. How else do you expect your employer to lay off your colleagues if you don’t assume their responsibilities by using AI tools?

This is why the primary challenge facing us in the AI world is not hallucinations, energy consumption, or bias, it’s how to care for people who might never be employed during their lifetimes. It’s also the reason that the tech bros are so hot for Universal Basic Income.

This raises existential questions for education. If schools are no longer to be factories for churning out workers, what is their mission in an AI world? Policy makers such as state education departments and local school boards can't delay starting to wrestle with these challenges. The world will be very different in a very few years. Today's kindergarteners deserve to be prepared for it starting right now.