7 things communicators need to know from Anthropic's 81,000-person AI study

Anthropic published what they believe is the largest qualitative study ever conducted - 81,000 interviews across 159 countries - asking people what they want from AI, what they've experienced, and what they're afraid of.

It's framed as a study about AI broadly. But read it through an internal comms lens and it's a study about your employees, your workplace and whether your organization is paying attention to any of this.

Here are the seven findings that internal comms practitioners should know.

1. Employees already made the decision
   Professional excellence was the number one thing people hoped AI would deliver — nearly 19% of respondents wanted it to clear away routine tasks so they could focus on higher-value work. And 81% said AI had already taken a real step toward their vision.
   
   Your employees aren't waiting for an AI strategy rollout. They're already using it, already benefiting from it, and already forming opinions about it. If your organization still hasn't communicated a clear position on AI in the workplace, you're not leading - you're catching up.
   

2. Time savings only count if someone actually gets the time
   Half of all respondents cited time savings as AI's biggest win. That's the good news.
   
   Here's the catch: 18% also described what the study calls "illusory productivity." The treadmill just speeds up. One freelancer put it plainly: the ratio of work time to rest time hadn't changed at all. "You just have to run faster and faster to stay in place."
   
   If your organization's AI pitch is built entirely on efficiency - doing more with the same people - employees will figure that out. And trust will take the hit. Time savings have to actually translate into something meaningful for the person doing the work, or the promise rings hollow.
   

3. Job fear is the most powerful force in the room
   Concern about jobs and the economy was the single strongest predictor of how people feel about AI overall. More than any other factor.
   
   That means IC can't soft-pedal this. Communications that highlight AI's upside while dancing around the job question aren't neutral. They're just unconvincing. Employees notice the gap. If leadership won't address displacement directly, IC should be advocating for that conversation to happen - not helping avoid it.
   

4. Forced adoption is a trust tax
   Nearly 22% of respondents named loss of autonomy and agency as a top concern. The study explicitly calls out forced AI adoption as one of the things people fear. One student described it this way: "The line isn't something I'm managing - it feels like Claude is drawing the line."
   
   Employees who feel like AI is being rolled out on top of them, not with them, will disengage. This isn't a change management problem you can fix with a FAQ page. It's a two-way communication problem. Are employees being asked what they think? Do they have real input, or just a feedback form that goes nowhere?
   

5. How you introduce AI changes what it does to people
   Educators in the study reported witnessing cognitive atrophy (skill loss from AI over-reliance) at two to three times the average rate, presumably in their students. Tradespeople learning on their own terms? Almost none, just 4%.
   
   The difference is choice. When people use AI because they want to, the benefits show up. When they use it because they're told to, shortcuts follow.
   
   The framing IC puts around AI tools is not cosmetic. It shapes how employees engage with them. A rollout that positions AI as "here to make your job easier" lands differently than one that positions it as "here to help you do better work." That's worth thinking carefully about.
   

6. Your employees can hold the complexity and your communications should too
   The study's defining finding is that hope and fear don't sort people into two camps. They live inside the same person. Someone who valued emotional support from AI was 3x more likely to also worry about becoming dependent on it.
   
   IC content that treats AI as either a win or a threat is less credible than the lived experience of the employees reading it. People are already sitting with both sides. Communications that acknowledge real concerns alongside real benefits will land. The ones that don't will get dismissed.
   

7. Your global workforce doesn't share the same AI story

   In wealthier regions, people wanted AI for life management — reducing cognitive overload and complexity. In developing regions, entrepreneurship and learning dominated. AI was described as a capital bypass mechanism, a way to build skills and businesses that would otherwise be out of reach.

   
   A single global AI story won't work. "AI will give you more time for what matters" hits completely differently for a knowledge worker in Copenhagen than for an entrepreneur in Uganda who might have their first real access to AI business tools. If your organization has a global footprint, this is worth building into how you segment and tailor AI communications.
   

The study is worth reading in full. The quote wall alone will give you more texture on employee sentiment than most internal surveys ever will.

The bigger takeaway for IC: this is 81,000 people telling us what employees actually think about AI, in their own words, without a comms filter on it. That's rare. Use it.

Want to listen a longer discussion? I discussed this study with Jenni Field on Frequency.

Written by Chuck Gose, founder of ICology.

Source: Huang, S. et al. (2026). "What 81,000 People Want from AI." Anthropic. https://anthropic.com/features/81k-interviews