Why AI productivity lags behind investments

While artificial intelligence is hailed as a revolution in efficiency, the productivity gains of AI remains massive despite companies pouring billions of dollars into experiments. This paradox reflects recurring patterns in the history of innovation. Groundbreaking technologies often trigger large investments long before they deliver measurable results.

With AI, companies are hiring talent, building tools, and testing automation strategies, but the productivity numbers are hardly moving. That's not necessarily a failure. It is a familiar “j-curve” and early efforts require large input before the conversion becomes entrenched. According to Stanford University's professor Jerry Yang and Yamazaki, the current issue is how to maximize the potential of AI, allowing businesses to move from costly pilots to truly reimagining. Senior Fellow at Stanford University's Human-Centered AI Institute. Director of Stanford Digital Economy Lab.

Erik Brynjolfsson, director of Stanford Digital Economy Lab, talks about the productivity impact of TheCube and AI and the practices needed to capture business value.

“By definition, it's not productive,” he said. “But when you understand that, things really take off. To me, it looks a bit like a J. [curve]and plot it. We wrote a paper on it in the American Journal of Economics. Now we are seeing the same thing as AI. ”

Brynjolfsson spoke with TheCube's Dave Velante and Rebecca Knight on Uipath Fusion during an exclusive broadcast on TheCube, the live streaming studio for Siliconangle Media. They discussed the impact of AI on productivity, the need for new metrics, and the management practices needed to acquire business value. (*Disclosure below.)

From pilots to AI productivity

According to Brynjolfsson, organizations deploying AI often default to incremental automation looking for cost savings rather than transformative changes. That approach reflects the early Internet age where companies locked themselves into older processes rather than reinventing technology. True AI productivity gains only emerge when you rethink how a company is designed and how a whole new business model will take shape.

“What I've been focusing on forever is to understand how we can turn these incredible technologies into business value, something I feel more than ever before,” he said. “We need to convert it into something that works for businesses, something that works for consumers. Unfortunately, it's not happening as quickly as possible.”

Scaling AI productivity beyond pilots requires more than enthusiasm. It requires creativity and patience. Imagine an entirely new kind of business that requires much more effort than automating a single task, but those who embrace reinvention can unlock much greater benefits, Brynjolfsson pointed out.

“I've always seen this and I'm trying to resist it,” he said. “In the end, it requires more creativity to imagine a whole new kind of business. It's not surprising that it often takes years to organize it. But we're starting to see some new kind of organization, some new business.”

Measuring the business value of AI

Traditional economic indicators miss much of the value digital technology creates, and the impact of AI is lacking, explained Brinjolfsson. However, new measures have gained the benefits of free or low-cost digital services. These metrics reveal how much AI is changing economic outcomes and help organizations target investments more effectively.

“The other thing we're working on is a new measure of the economy,” he said. “We've developed this scale called GDP-B, which measures the value created without paying. As the economy gets more and more digital, there's a huge amount of value that doesn't appear anywhere else in the statistics. We're trying to measure it.”

Despite the pain of short-term growth, Brynjolfsson envisions a future in which AI productivity accelerates as intelligent systems enhance the human mind.

“If AI really recognizes the promise to enhance intelligence and automate it, it may be literally the last invention we have to make, because it starts inventing something new,” he said. “It starts to boost R&D, and I just read about regular AI, which is the company, it boosts scientific discovery.

This is a full video interview that is part of Siliconangle and part of The Cube's Uipath Fusion coverage.

https://www.youtube.com/watch?v=sddy5mcila8

(*Disclosure: TheCube is a paid media partner of UIPATH Fusion. Neither Uipath Inc., sponsor of TheCube's event coverage, nor can edit content from Thecube or Siliconangle.)

Photo: Silicon Angle