rewrite this title Chart of the Week: The AI Adoption Gap

This week, I’ve shown you how artificial intelligence is racing toward artificial general intelligence (AGI).

It can solve elite math problems, conduct scientific research, write sophisticated software and work independently for longer than ever before.

Yet for all of those breakthroughs, most businesses still aren’t using AI anywhere near its full potential.

This week’s chart helps explain why.

Untapped Opportunities

Earlier this year, Anthropic published a research paper that may explain one of the biggest mysteries in artificial intelligence.

If today’s AI is becoming so capable, why aren’t businesses seeing bigger results?

The company compared what today’s AI models could theoretically do across dozens of occupations with what people are actually using AI to do in their daily work.

And the difference is astonishing.

But you need to understand what this chart is actually measuring before you can understand the impact of this study.

Let me explain.

Most studies estimate AI’s potential by examining the tasks that make up different occupations and asking how many of those tasks today’s models could theoretically perform.

Anthropic did that too. But then the company took an extra step.

It analyzed millions of conversations people had with Claude to estimate how often workers are actually using AI for those same tasks.

Anthropic calls this “observed exposure.”

In other words, instead of measuring AI’s potential, it measured AI’s adoption.

The blue area shows AI’s theoretical exposure, while the red area shows its observed exposure. And the gap between the two shows us just how much AI capability businesses are leaving on the table today.

Look at management, business and finance, legal and computer programming. Across nearly every white-collar profession, the blue area towers over the red.

That tells us today’s AI appears capable of contributing to far more work than it’s currently being used for.

Computer and math occupations are an obvious example. Anthropic estimates today’s AI could theoretically cover roughly 94% of the tasks performed in those jobs.

Yet the observed usage is only about one-third of that.

The pattern repeats throughout much of the chart. Even in occupations like arts and media and life and social sciences, the gap is quite large.

To me, this shows that adoption is the bigger bottleneck than intelligence.

But that shouldn’t come as a complete surprise to you.

As I wrote about yesterday, businesses aren’t looking for AI that occasionally produces brilliant work. They need systems they can rely on day after day that fit into existing workflows, comply with regulations, integrate with other software and ultimately save more money than they cost.

That’s a high bar.

Anthropic points to several reasons this adoption gap still exists, including technical limitations, legal restrictions and the need for human oversight.

I’d add something else.

Most companies are still learning where AI actually creates value.

I’m convinced that’s why so many CEOs remain disappointed with their early AI investments. It’s also why Wall Street has become increasingly focused on AI revenues instead of benchmark scores.

To me, that’s less a reflection of AI’s capabilities than a reminder of how innovation spreads.

The thing is, most businesses don’t usually adopt new technologies overnight.

As I mentioned yesterday, the automobile didn’t transform America the day the first car was built. And the internet didn’t reshape commerce from the moment people first logged online.

Both required years of infrastructure buildouts, new business models and changing habits before they fulfilled their potential.

That’s where the AI industry finds itself right now.

It’s ready to transform the world, but it’s still earning the trust of the businesses it’s meant to help.

Here’s My Take

As we inch closer to artificial general intelligence (AGI), businesses are discovering that intelligence alone isn’t enough.

Today’s chart helps reconcile those two ideas.

Tomorrow, I’ll show you how the solution to this problem could come from AI itself.

Regards,

Ian KingChief Strategist, Banyan Hill Publishing

Editor’s Note: We’d love to hear from you!

If you want to share your thoughts or suggestions about the Daily Disruptor, or if there are any specific topics you’d like us to cover, just send an email to dailydisruptor@banyanhill.com.

Don’t worry, we won’t reveal your full name in the event we publish a response. So feel free to comment away!