91% of Anthropic staff expect a Nobel-level AI discovery by 2030, but 39% of Americans do

Asked to forecast unemployment, Americans and Anthropic employees reason the same way, and only their sense of what AI will do differs.

By Think Facility · October 17, 2026 · Anthropic AI policy

On October 16, 2026, Anthropic published a survey comparing what its own staff and ordinary Americans expect AI to do by 2030. The gap is wide. 91% of Anthropic staff expect AI to be able to make a Nobel-level discovery by 2030, against 39% of US adults.

Morning Consult surveyed 10,980 US adults (weighted to match the adult population) between August 11 and 23, and 1,334 Anthropic employees answered an internal poll from August 9 to 12. The report is by Szymon Sacher, Tess Cotter, Charles I. Jones, Anton Korinek and Peter McCrory, and it came out alongside Anthropic's report on how its own hiring has changed.

Nobel-level discovery by 2030: 91% of staff, 39% of US adults
Own work automatable, 2030: 75% for staff, 28% for US workers
Task in quarter time: 59% of staff, 26% of US adults who use AI at work
Implied 2030 GDP growth: 17.1% for staff, 5.4% for US adults

Where the two groups split

The two groups mostly agree on what AI can do today, and they rank tasks in nearly the same order of difficulty. They split on the future. Depending on the task, between 19 and 40% of US adults say AI will never be able to do it. At most 2% of staff say that about any task.

The gap is sharpest on people's own jobs. Staff expect AI to be able to do 75% of their current work hours by 2030, on average. US workers put it at 28%, and only 1 in 20 expect 80% or more.

Use explains part of it. Among Americans who use AI at work regularly (about a fifth of the sample), 56% expect it to manage at least seven of the eight named tasks by 2030. That's 41% for all US adults and 28% for people who've never used an AI tool. Staff are at 96%.

Most Anthropic staff already work alongside AI that does much of their job, whereas few US workers do.

From The AI expectations gap \ Anthropic

What the two groups agree on

Here's the finding I'd put first. People were told unemployment was about 4.2% and asked to forecast it for 2030. In both groups, the more of their own work someone expects AI to do, the higher the unemployment they forecast. US workers who expect AI to do 80% or more of their work forecast 12.2%, which the authors note is above the peak during the Great Financial Crisis.

At the same level of expected automation, staff and the public forecast about the same rate. So the authors put the gap down to experience (staff use AI all day, and most Americans don't). On how jobs work, the two groups think much alike.

Staff also expect Anthropic to keep hiring. 95% expect headcount to grow over the next 12 months, and the median guess is 66%. The report calls these personal views from an internal poll and says they don't represent Anthropic's plans.

The number about AI building AI

One question went only to staff. It asked what share of the work of building an early-2026 frontier model AI does today. The median answer was 51%, rising to 80% by the end of 2027 and 95% by the end of 2030. Research staff, closest to the work, put today's share at 60% on average.

The report says plainly that these are personal estimates. Still, it's the people who build the models saying AI already does about half of that work, and I'd read it with that in mind.

What it means for jobs and the economy

The 4.2% people were shown is where the country actually sits. The Bureau of Labor Statistics reported on October 2, 2026 that unemployment was 4.2% in September, with 7.1 million people out of work, and that the rate has stayed between 4.1% and 4.3% since March. It's a long way from 12.2%.

Both groups also expect it to be hard to change careers. The median staff member thinks a worker who loses a job to AI in 2030 and has to switch occupations would take 12 months to find one. Most Americans picked 7 to 12 months. Both were read the same baseline first: in normal times, about 3 months. So even the public expects a displaced worker to wait two to four times as long.

The growth gap is the other half. Run through Anthropic's own economic model, the median American's answers imply 5.4% GDP growth in 2030, which the authors say is more than twice the typical rate. The median staff member's answers imply 17.1%, an economy doubling roughly every four and a half years. 23,075 visitors who took a short version of the quiz on Anthropic's site landed in between, at 12.5%.

Public views are moving a little. When Morning Consult asked again from September 17 to 24, the share of US adults expecting a Nobel-level AI discovery by 2030 rose to 44%. The BLS publishes October's jobs numbers on November 6, 2026.