The AI Awakening: Insights from the Business & Generative AI Conference

At the 2024 Business & Generative AI Conference, hosted by Wharton Human-AI Research, Stanford’s Erik Brynjolfsson delivered a keynote that challenged business leaders to look beyond the headlines and toward the deeper economic transformations generative AI is already driving.

A pioneer in digital economics and Director of the Digital Economy Lab at Stanford HAI, Brynjolfsson made a compelling case that AI represents a foundational shift in how we work, who benefits, and how prosperity is distributed.

“AI is the most general of all general-purpose technologies,” he said. “It’s changing what we can do, how we work, and who benefits.”

Here are four key takeaways from his talk:

Traditional Metrics Won’t Capture the AI Economy

Brynjolfsson emphasized that as AI systems become more capable, our current economic indicators will increasingly fail to reflect what’s really happening. Measures like GDP and labor productivity, he argued, were designed for an industrial economy, not one driven by digital platforms, open-source tools, and distributed intelligence.

“The tools we use to measure progress were built for the last revolution. They’re not ready for what comes next,” he said.

For example, AI can dramatically improve decision-making, customer experience, or time-to-market, yet these gains often don’t show up in GDP because they involve quality improvements rather than quantity of output.

Brynjolfsson urged business and policy leaders to develop new ways of tracking value in an AI-driven economy – from more nuanced productivity metrics to broader wellbeing indicators – so we can better guide investment, education, and innovation.

AI as a Productivity Multiplier (Especially for the Underserved)

In a large-scale field study with over 5,000 call center agents, Brynjolfsson and his team found that generative AI tools improved productivity by 14% and had the most impact on the least experienced workers, who improved by up to 35%.

“This wasn’t skill-biased technical change. It was the reverse. The AI helped people with less experience catch up.”

Key results:

• Faster resolutions and higher customer satisfaction
• More consistent and effective responses across teams
• Accelerated skill development through real-time feedback

The AI system acted as a real-time coach, distributing tacit knowledge such as tone, phrasing, and effective response strategies that previously only experienced workers possessed. Brynjolfsson emphasized that this kind of augmentation has broad implications for frontline teams, onboarding, and workforce development, especially in sectors where experience gaps are common.

Escaping the “Turing Trap”: Why Augmentation Beats Automation

While some technologists pursue human-like AI as the end goal, Brynjolfsson urges a shift in mindset: stop trying to replicate humans and start building systems that empower them.

This “Turing Trap,” as he calls it, is the tendency to design AI to mimic human workers, aiming for full automation. This approach, he warned, risks concentrating power and eliminating labor income altogether, even as productivity soars.

Instead, Brynjolfsson called for a deliberate focus on augmentation: designing AI to complement human strengths, expand capabilities, and support shared prosperity.

Preparing for Transformative AI: From Forecasts to Frameworks

Brynjolfsson believes we are rapidly approaching a new threshold in AI development, one he refers to as Transformative AI: systems with the capacity to affect nearly every process, role, and industry. This shift is not just about replacing tasks, he emphasized, but about redefining the structure of the economy itself.

To help guide that transformation, Brynjolfsson introduced a new research agenda, Project Apollo.

Named for the scale and urgency of the original moonshot effort, Project Apollo seeks to explore the institutional, economic, and policy innovations required in a world increasingly shaped by intelligent systems.

Key areas of focus include:

• New productivity benchmarks: How do we measure output and economic value when humans and AI work together in creative, non-linear ways?
• Labor and income structures: What happens to employment, wages, and social mobility if AI systems increasingly perform high-value knowledge work?
• Policy and institutional design: What frameworks can help ensure that prosperity is widely shared, rather than concentrated among a few owners of capital and technology?

Why It Matters

Brynjolfsson’s talk aligned with the Business & Generative AI Conference’s central mission: not just to explore what AI can do, but to shape how it is applied, with a focus on responsible leadership, real-world impact, and inclusive innovation.

His research underscores that the biggest gains from generative AI will come not from replacing workers, but from empowering them, and from building the business and economic structures needed to support that vision at scale.

Register for the 2025 Business & Generative AI Conference here.