The Skills You Need in the Age of AI

Imagination is more important than knowledge.
Albert Einstein

Artificial intelligence is having its moment. If you even casually follow trends in business and productivity, it’s hard to escape the predictions that AI will soon take over the world, and perhaps your job.

Should you be worried that AI will replace you?

Perhaps.

It’s capabilities are robust. It can synthesize vast troves of information, summarize books, give you what you need in an instant. It can generate images and videos, answer questions about seemingly anything, help you plan your day, schedule meetings, and more.

If you trade in knowledge and basic analysis — regurgitating facts and figures, curating, reassembling, and conveying information, running numbers and financial models — you might want to reconsider your work.

This is the time to look beneath the labels of what you do and consider the broader scope of your work. What do you really do?

On the other hand, if you work with people in a meaningful, service-oriented way, or if you are a true creative, you have little to fear.

The Limits of AI

In a previous essay, I wrote about the limits of labels. In brief, the language we use to define something can be a helpful shorthand that makes communication more effective. At the same time, labeling something can limit its potential.

I used the example of a table to illustrate my point. If you didn’t know what the object was, you might approach it with a different perspective, unlocking its potential to be used in different ways.

To illustrate some of these possibilities, I asked AI to create an image of a table turned upside down and reimagined as a playpen. I described exactly what I wanted with detail. AI failed to generate it.

I revised the prompts and tried again. AI still couldn’t get it.

After a few rounds, I asked AI to help me craft better prompts. It accurately summarized what I was requesting, showing that it clearly understood the concept. But even with its own prompts, it couldn’t generate the image.

Finally, I asked AI why it couldn’t seem to generate the image I wanted. It gave me a lengthy response, but here’s the essence:

The challenges you’ve experienced seem to stem from trying to accurately capture a very specific and creative concept that involves a nuanced understanding of the objects and their intended use. While the system strives to follow the provided instructions as closely as possible, achieving the exact vision described can sometimes be outside its current capabilities, especially when the request involves unconventional uses or arrangements of objects that are not commonly depicted together in its training data.

AI couldn’t create the image I wanted, but it inadvertently proved the point of my essay: when we define or label something, we limit its potential.

AI can’t see beyond current definitions and established norms. It cannot listen to what’s beneath the surface of what someone says.

In effect, AI can only create something that is derivative of what humans have already created. It can regurgitate on a level of sophistication that humans cannot match, and it can possibly pull together disparate ideas.

The tool of AI might be innovative in terms of how we use it in our work, but AI itself cannot innovate. It can combine things, but it can’t truly be creative.

The tool of AI might be innovative in terms of how we use it in our work, but AI itself cannot innovate. It can combine things, but it can’t truly be creative.

AI can give us access to knowledge and help us better understand things, but it cannot imagine what doesn’t exist.

We still need humans for that.