With ChatGPT surpassing one billion users in 2025, nearly everyone uses AI to some capacity. American workers aren’t exempt. They’re not waiting for permission, either. Instead, they’re quietly revolutionizing their jobs from the inside out. To understand how workers are using AI, we surveyed U.S. employees across a wide range of sectors and industries. From secret AI adoption to anxiety and compensation concerns, the data uncovers a workforce navigating uncharted territory with little guidance. Employers may not realize how quickly AI is transforming the workplace. Keep reading to learn more.
Key takeaways
Four in five people use AI at work, and 36% consider it “essential” to their job.
66% of workers are personally funding AI tools for work.
45% of employees have used AI at work without informing their manager, with Gen Z and tech workers being the highest demographic to do so.
Nearly half of workers (47%) have taken actions to safeguard their professional and financial futures due to AI uncertainty, including increased savings (34%) and starting side hustles (24%).
The secret AI revolution: workers go incognito
While executives debate AI policies in boardrooms, their employees may already be deciding for them. Our survey reveals a hidden revolution taking place across American offices, where workers are quietly integrating AI into their daily routines—often without their managers’ knowledge.
Interactive or visual content
The numbers tell a striking story. Nearly half of American workers (45%) have used AI tools at work without informing their managers or employers. This underground adoption is happening even as only 26% of workers report having clear company AI policies.
The implications are profound. AI adoption is outpacing organizational readiness by a significant margin, creating a shadow economy of productivity gains that employers may not even realize they’re benefiting from. Workers are taking initiative and figuring it out as they go, fundamentally reshaping how work gets done in America.
The AI skills anxiety: 1 in 4 workers overstate capabilities
The pressure to demonstrate AI competency has created a job market where authenticity takes a backseat to survival. Our data reveals that a significant portion of American workers are gambling with their careers, betting they can learn AI skills fast enough to justify their claims.
Interactive or visual content
The combined data reveals that one in four workers (25%) have either exaggerated their AI capabilities in job applications or intentionally overstated their skills during the hiring process. This includes 13% who admitted they “figured it out later” and 13% who “have since caught up” to their initial claims.
This widespread skills inflation reflects the intense competitive pressure surrounding AI literacy in today’s job market. Workers understand that AI competency is rapidly becoming table stakes for career advancement, but formal training opportunities haven’t kept pace with demand. The result is a “fake it ’til you make it” culture where career survival often depends on bold promises about future capabilities rather than current expertise.
The AI job anxiety map: state-by-state search trends
While AI is becoming more popular overall, its popularity varies from state to state, and the same holds true for anxiety surrounding the new technology. Our analysis of Google Trends data for 100 AI job-related anxiety terms reveals striking differences in concern levels by state, spotlighting where the conversation is most intense.
Interactive or visual content
Our findings show that California, Florida, and Texas top the nation for total AI job anxiety search volume, reflecting both their large workforces and significant exposure to technology-driven change. When adjusted for population, smaller states like Colorado and Mississippi reveal even higher rates of AI job anxiety per capita.
The most commonly searched concern in high-anxiety states centers on “AI replacing jobs” and the broader “technology job market,” underscoring both specific and general fears about the future of work. States with higher AI anxiety may have more workers who worry about the longevity of their career, and whether they may need to branch out to a position that AI won’t replace. This geographic breakdown highlights that AI job anxiety is not just a coastal or tech-hub phenomenon—it’s a nationwide issue demanding attention at every level.
The AI dependency divide: essential vs. experimental usage
Despite the growing popularity of AI, many American workplaces are holding fast and refusing to integrate it into their workflows. These workplaces stand in stark contrast to their counterparts, which integrate AI into every aspect of their business, from helping to brainstorm new ideas to handling more administrative tasks. This divide reveals fundamental differences in how workers view AI’s role in their professional future.
Interactive or visual content
Our data also corroborates this divide: more than one-third of workers (36%) now consider AI essential for their daily tasks, while the majority still view it as supplementary or optional. This dependency gap has significant implications for workplace productivity and competitive advantage.
Perhaps most telling, 53% of workers say their productivity would slow down if AI were banned tomorrow, indicating that AI has already become embedded in American work infrastructure. This isn’t just about convenience. It represents a fundamental shift in how work gets done, with early adopters gaining competitive advantages while others risk falling behind.
How workers disclose AI usage
Even though many U.S. employees recognize the benefits of AI, a majority of workers still hesitate to disclose their AI use, especially to their employers. Our survey reveals a significant transparency gap, with most workers opting for ambiguity over clarity when presenting AI-assisted work to colleagues and supervisors.
Interactive or visual content
The transparency data reveals that a workplace culture is still grappling with the ethics of AI integration. While nearly one-third of workers (32%) proactively explain their AI usage, the majority take a more ambiguous approach—either providing minimal detail, adjusting their disclosure based on context, or avoiding mention entirely.
This communication gap has significant implications for workplace trust, collaboration, and quality assessment. When 69% of workers aren’t fully transparent about AI assistance, it becomes difficult for managers to accurately evaluate work quality, provide appropriate feedback, or make informed decisions about task assignments and skill development.
Who gains from AI?
A significant question workers may ask when considering the use of AI is, “How will this improve my daily life?” The answer may vary, and to some degree, AI may not even improve the lives of workers directly, but rather the company they work for. In fact, 39% of corporations have already experienced a moderate increase in ROI due to the use of AI in 2025. Instead of distributing their monetary profits to employees who have trained themselves on effective AI prompts, employees worry that their employers may retain the financial compensation and use it for corporate growth. Not only does this impact employees’ attitudes toward AI, but it may also affect their perception of their companies, their positions, and their careers as a whole. Our survey reveals a growing perception gap between AI-driven company profits and individual worker compensation.
Interactive or visual content
Nearly a quarter of workers (24%) believe their companies are profiting from the implementation of AI, while they see no personal financial benefit. An additional 17% report being expected to do more work without further compensation.
This perception of unequal benefit distribution could have long-term implications for employee satisfaction and retention. When workers feel they’re driving AI-powered productivity gains without sharing in the rewards, it creates potential tension between individuals and their employers. Only 17% of workers report receiving tangible rewards, such as raises or bonuses, tied to AI-enhanced performance.
Embracing the AI shift
Our study reveals a fundamental shift in how workplace technology adoption occurs — many workers are no longer waiting for organizational permission to integrate AI into their daily routines. This bottom-up transformation represents a new paradigm where individual initiative drives technological change faster than corporate policies can accommodate. The implications extend far beyond productivity metrics. Instead, it may have a significant impact on how employers and employees handle the implicit contract of AI use, which begins as early as onboarding.
For business leaders and HR professionals, this worker-led AI revolution demands a strategic response that acknowledges current realities rather than attempting to control an already established trend. Companies that adopt AI while establishing clear guidelines for responsible use will likely see greater employee engagement and competitive advantages, especially as 53% of workers report that their productivity would slow down if AI were banned.
The path forward requires striking a balance between employee autonomy and organizational oversight, creating frameworks that legitimize AI usage while maintaining standards for transparency and accountability, and developing compensation structures that recognize the contributions of AI-enhanced productivity.
Methodology
We surveyed 1,000 U.S. workers in June 2025. The sample was stratified by age, gender, industry, and region to ensure balanced representation across key demographics. We also analyzed Google Trends data for 100 AI job anxiety-related terms across U.S. states from June 2024 to June 2025. This approach allowed us to capture both self-reported workplace sentiment and real-time geographic patterns in AI job anxiety, providing a comprehensive view of how Americans are experiencing and searching for AI’s impact on jobs.
Fair use statement
This content is available for non-commercial use. When referencing this study, please include a link to this page.
