Fed's Mary Daly Observes Early AI Productivity Gains in Select Sectors

Mary Daly, President and CEO of the San Francisco Federal Reserve, says artificial intelligence is starting to show real productivity improvements at specific companies and in certain sectors. The catch: those gains haven’t shown up in the broader economy yet.

Productivity growth has climbed from a historical average of roughly 1.9% annually to 2.7% over the most recent 10 quarters. That’s a meaningful jump. But Daly was careful to note that it’s difficult to determine how much of that improvement is specifically attributable to AI versus other factors.

Daly first laid out her framework on AI and productivity in remarks delivered on February 17, 2026. At that time, she noted that most macro-studies showed limited evidence of a significant AI effect on the economy. In more recent comments from late May and early June 2026, she described what she sees as “green shoots,” acknowledging firm-level improvements while maintaining that the data on economy-wide productivity gains remains thin.

Think of it like electricity in the early 1900s. Factories had to be completely redesigned before electrification delivered its full productivity punch. The technology existed for years before the economic statistics reflected its impact. Daly has drawn this parallel explicitly, suggesting AI may follow a similar adoption curve.

One of the more underappreciated parts of Daly’s commentary involves the obstacles standing between AI adoption and measurable productivity gains. She identified regulatory and legal barriers as significant hurdles that industries face when trying to leverage AI effectively.

Despite historically high levels of investment enthusiasm around AI, the data doesn’t yet show the kind of widespread productivity benefits that would move the needle on overall economic trends.

The core scenario Daly is sketching looks like this: if AI can deliver durable productivity increases, economic output could grow faster without necessarily stoking inflation. She has suggested the upcoming year could serve as a critical period for evaluating whether AI’s economic effects are becoming more tangible.