top of page

Stanford AI Worker Survey: Key Findings

The Stanford AI Worker Survey, which is part of the broader Stanford AI Index 2025 Annual Report, was officially released in April 2025 by the Stanford Institute for Human-Centred AI. This annual report provides comprehensive insights into the global state of artificial intelligence, including the latest findings on AI's impact on the workforce and the future of work.


Workers Don't Want AI to Replace Them — They Want AI as Their Partner

What if everything we think we know about AI and the future of work is wrong? Recent research from Stanford suggests that it might be.

 

The Stanford Institute for Human-Centred AI has just released findings from their comprehensive AI Worker Survey, which is part of the Stanford AI Index 2025 Annual Report. They surveyed 1,500 workers across 104 occupations to understand what people actually want from AI in the workplace, and the results are surprising.


The Partnership Revolution

Here's the headline finding that should make every tech company pause workers don't want AI to replace them; they want AI as an equal partner. The research team from Stanford's Digital Economy Lab discovered that the most commonly desired arrangement is "equal partnership" between humans and AI. This preference emerged as the dominant choice in 47 out of the 104 occupations studied.

 

To measure this, Stanford created the Human Agency Scale (HAS), a five-level metric that quantifies how much human involvement workers want in their tasks. The scale ranges from full automation (H1) to human-essential (H5). The findings reveal that workers generally prefer higher levels of human agency than what AI experts believe is technologically necessary for 47.5% of tasks.


What Workers Really Fear (And It's Not What You Think)

When examining worker concerns about AI, job replacement ranked second. The biggest worry? Trust.


Here's what keeps workers awake at night about AI:

  • Lack of trust in AI systems: 45%

  • Fear of job replacement: 23%

  • Loss of the human touch: 16.3%


This highlights something crucial: the trust issue is more significant than the job replacement issue.


When Workers DO Want Automation

Interestingly, workers are not opposed to automation across the board. When they want AI to take over tasks, here's why:

  • Freeing up time for high-value work: 69.4%

  • Handling repetitive tasks: 46.6%

  • Improving quality: 46.6%

  • Reducing stress: 25.5%


The pattern is clear: workers want AI to manage the boring, repetitive, and stressful aspects of their jobs so they can focus on more meaningful work.


The Four Zones of AI Adoption

Stanford's researchers identified four distinct zones where AI fits into the workplace:

  1. The "Green Light" Zone: High worker desire meets high AI capability. These are the sweet spots where AI deployment can deliver real productivity gains and worker satisfaction.

  2. The "Red Light" Zone: High AI capability but low worker desire. These areas require careful handling because they may face resistance or create negative social impacts.

  3. The R&D Opportunity Zone: High worker desire, but current AI capabilities aren't sufficient. These indicate promising directions for future AI research.

  4. The Low Priority Zone: Low desire and low capability, meaning they are simply not worth pursuing right now.


A Skills Revolution is Coming

As AI becomes more capable, the job market is heading toward a major shift. The demand for information-processing skills is expected to shrink, while interpersonal and organisational skills will become increasingly valuable. Future skills like decision-making, coordination, and quality judgement will take precedence over data analysis. Workers who can manage relationships, train others, and navigate complex human interactions will be in high demand.


The Investment Mismatch

This poses a concerning dilemma for the tech industry: there is a significant mismatch between what workers want and where AI investment is currently flowing. Stanford found that 41% of Y Combinator company-task mappings are concentrated in areas where workers have a low desire for automation. The industry is building solutions for problems that workers don't actually want solved, while neglecting the areas where they are asking for help.


What This Means for the Future

Stanford's research presents a future that is fundamentally different from the typical "AI takeover" narrative we often hear. Workers are not asking to be replaced; they are asking for better tools and smarter partnerships. The study suggests that successful AI adoption will require acknowledgement of varying levels of desired human agency. Companies that ignore worker preferences may face resistance, while those that embrace a partnership model could unlock significant productivity gains.

 

The future of work isn't about humans versus AI. It's about humans collaborating with AI in ways that respect both human agency and technological capability. The workers have spoken — now it's time for the tech industry to listen.

 

This research is part of the Stanford AI Index 2025 Annual Report, conducted by the Stanford Digital Economy Lab and its collaborators. Full findings and methodology are available at futureofwork.saltlab.stanford.edu.



Stanford findings and why AI-CAE an excellent next step for leaders in associations and membership bodies:


1. Addressing the Trust Gap (45% Worker Concern) AI-CAE includes a core module on "AI Governance – Build trust, control risk, lead responsibly." Since the Stanford research shows that the lack of trust in AI systems is the biggest concern for workers at 45%, association leaders need specific governance frameworks to build this trust among their members and staff.


2. Partnership-Focused Approach Aligns with Worker Preferences The Stanford research reveals that workers generally prefer higher levels of human agency than what AI experts believe is technologically necessary for 47.5% of tasks. AI-CAE is described as "not just another AI course," but instead focuses on practical leadership that supports the partnership model workers actually want.


3. Avoiding the Investment Mismatch Problem Stanford found that 41% of Y Combinator company-task mappings are concentrated in areas where workers show low desire for automation. AI-CAE's module on "AI Use Cases – Spot high-impact opportunities across your organisation" specifically helps leaders identify the "Green Light" zones where both worker desire and AI capability are strong.


4. Member Engagement in the Skills Revolution The research shows that interpersonal and organisational skills will become increasingly valuable. AI-CAE's module on "AI & Member Engagement – Deliver personal, data-driven experiences" assists association leaders in leveraging AI to enhance, rather than replace, the human connections that will become premium skills.


5. Addressing the 72.7% Knowledge Gap AI-CAE addresses the fact that "72.7% of association executives say they lack the knowledge to lead AI adoption," and "Only 1 in 5 have a strategy." This directly connects to Stanford's findings about the need for leaders who understand what workers actually want from AI partnerships.


6. ROI Aligned with Worker Motivations Stanford found that workers want AI to free up time for high-value work (69.4%) and handle repetitive tasks (46.6%). AI-CAE promises to "Increase member retention by 18%" and "Reduce response times by 60%"—outcomes that align with using AI for efficiency rather than replacement.


7. Culture and Change Management for Partnership Model AI-CAE includes a module on "Culture & Change – Win hearts, build confidence, support staff," directly addressing Stanford's finding that successful AI adoption requires acknowledgement of varying levels of desired human agency.


8. Strategic Focus Over Technical Focus AI-CAE emphasises that "You don't need to be a tech expert—just a leader with a plan," and includes testimonials stating, "It's not about understanding the tech—it's about leading strategically." This perfectly aligns with Stanford's findings that the challenge lies not in technical capability but in understanding worker preferences and building trust.


The AI-CAE programme is specifically designed to help association leaders navigate the exact challenges identified in Stanford's research: building trust, focusing on partnership rather than replacement, and developing strategies that align with what workers actually want from AI integration.


 

 

 

 
 
 
bottom of page