Absolutely formidable: Google’s chief economist on the impact of AI

"Remember, I am sitting in the chocolate factory. So I see a little bit further what we've got coming down the lane, and it is absolutely formidable."

Fabien Curto Millet, the chief economist at Google, has a ring-side seat on the accelerating developments in AI, and spends his days working out the economic impact of the technology. So the Forum podcast Radio Davos wanted his reflections as we publish the latest Chief Economists Outlook - an overview of where we might expect the global economy to head.

You can watch the video-podcast here:

Or get the audio on any podcast app, via this link, or listen here:

Get it on any podcast app via this link: https://pod.link/1504682164

Here are some highlights of the interview:

“My job is essentially a mixture of four things,” said Millet.

“Firstly, I need to grapple with the global environment Google operates in. Secondly, how we use economic tools to improve our business decisions within this reality. Thirdly, how do we navigate public policy shaping the tech sector, especially antitrust. As a large company, we’re always in the spotlight in multiple jurisdictions. Finally, thought leadership – as AI reshapes society and the economy, what role can Google play in responsibly guiding this transformation?”

The Chief Economists Outlook highlights concerns about geopolitical instability and rising economic nationalism fuelling global uncertainty.

Millet gave said organizations should understand and prepare for different kinds of uncertainty.

“Economists distinguish between decision-making under risk and decision-making under uncertainty. The former deals with knowing the probabilities of the decision, and the latter with completely uncertain realities.”

Decision-making under uncertainty is, according to him, “Like playing dice without knowing how many faces it has—or what’s written on them!”

Millet says proper scenario planning is the way to prepare. By reducing uncertainties into manageable risks, an organization is empowered to develop strategies that consider most potential outcomes.

Millet describes AI as “the most exciting technological shift” of his lifetime and says that the evidence of its impact is prevalent across most aspects of our global economy.

“On a grassroots level, AI’s impact is already quite clear,” he said. “AI boosts software developer efficiency by 21%, speeds up professional writing by 40%, and increases call centre productivity by 14%.”

Right now, AI adoption is still low—under 6% of American firms use it in production. Goldman Sachs, however, has forecast that AI could lead to a 15% increase in the GDP of the USA, while the OECD predicts a 10% gain over the next decade, Millet says.

What about the uncertainties over these outcomes? Millet said any concerns about a hype cycle like the ‘dot-com bubble' (late 1990s – early 2000s), should be offset by the longer term promise of AI.

“Yes, enthusiasm during the dot-com era led to huge investor speculation and financial overreach. However, the ultimate transformative impact of the internet age is quite apparent today. Despite that period, the entire economy today has been reshaped by digital markets, which were in their infancy 30 years ago,” he said. “AI’s impact will be faster and deeper. This is why we need to responsibly prepare for its risks.”

What about the concern that AI may increase the inequalities between developed and developing economies?

“I don’t agree with this narrative. Surveys show more enthusiasm for AI in the developing world than in the advanced world. At Google, we see building out our digital infrastructure as a way to enable more people in emerging markets to participate. Our subsea cables, for example, are having a major impact on West African communities’ ability to connect.”

He added: “In a place like Togo, for example, where the median age is 19 and teachers are scarce, AI is proving to be a lifeline for education. It’s empowering teachers and expanding access to quality resources.”

The massive demand for new data centres to power AI means the energy required is skyrocketing. Millet, says better, more efficient technologies will help soften that blow.

“What many don’t realize is that data centers, today, are 4x more compute-efficient than they were 5 years ago. Our latest Tensor Processing Units [Google’s custom-developed, application-specific integrated circuits used to accelerate machine learning workloads] are 30x more efficient than the first generation. That is a fantastic testament to innovation.”

He admits, however, that energy remains a major challenge. “The U.S. is projected to need 128 additional gigawatts by 2030. For its part, Google is investing in things like advanced geothermal and modular nuclear power, grid-enhancing technologies, and training new electricians.”

One of the biggest fears about AI is mass unemployment. Millet counters that “Technology replaces tasks, not jobs. Most jobs are made up of multiple tasks, many of which are hard to automate economically.”

“Employees are adopting AI before employers even realize it. A McKinsey report showed companies underestimate AI usage internally by a factor of three. Historically, we’ve seen huge shifts—like agriculture dropping from 60% to under 5% of employment in the USA without mass joblessness.”

He likens the current moment to the early days of electrification or computing: Transformative technologies that took time to reorganize the economy around them.

“With the right policy,” he emphasized, “AI can create more opportunities than it displaces.”

“I am truly amazed at what Gemini (Google’s AI) can accomplish. Recently, it helped a team at Imperial College generate a hypothesis about antibiotic resistance in minutes – that had taken them ten years to develop!”

For Millet, the current moment is pivotal for humanity.

“We live in a time of wonders,” he said. “Quantum computing, AI, climate solutions, revolutionizing healthcare and education —all within reach. It’s up to us to design the future we want. It’s a challenge – but an opportunity beyond our wildest comprehension.”