Generative artificial intelligence has gone from being an occasional tool to becoming part of the daily routine for millions of workers. Half of the workers surveyed by ADP Research use applications like ChatGPT, Copilot, or Midjourney several times a week, while 20% say they use them almost every day. However, this adoption doesn’t necessarily translate into a greater sense of productivity.
The key points of AI use at work in 30 seconds
- 50% of workers use AI several times a week, and 20% do so almost daily.
- Daily users show more engagement and less negative stress than those who never use these tools.
- They are also four times more likely to recognize that they could be more productive.
- In Spain, 15% use AI almost every day, and 17% have not tried it yet.
- ADP highlights training and workflow adaptation as areas for improvement for companies.
The data comes from People at Work 2026, an ADP Research report based on responses from more than 39,000 adult workers across 36 markets. The survey was conducted between July 21 and August 4, 2025, with samples organized by age and gender in each country.
The study does not directly measure how many tasks each employee completes or how much time they save with AI. It analyzes workers’ perceptions of their employment, productivity, stress, job security, and relationships with their teams. Therefore, its results describe perceptions and associations, but do not establish that artificial intelligence by itself causes the observed changes.
More engagement and less stress among AI users
Frequent use of artificial intelligence is associated with a more positive work experience across several indicators.
30% of those who use AI almost every day consider themselves fully engaged with their work. Among those who never use it, the proportion drops to 14%. As a reference, full engagement among the global workforce stood at 19% in 2025.
The difference also appears when analyzing stress levels. Only 11% of daily users were classified as overburdened workers, compared to 23% among those who had never used AI.
ADP differentiates between positive stress, which can keep a person active and motivated, and negative stress, linked to exhaustion and feeling drained. Regular AI users reported fewer experiences of this negative type.
The research also finds a relationship between tool usage and team perception. Those who use AI more frequently are more likely to say they are part of a team and, especially, to feel they belong to the best team in their organization.
These findings do not prove that installing an AI assistant will automatically improve the work environment. Employees who are more engaged, better resourced, or part of tech-savvy teams may also have greater access to these solutions.
Nonetheless, the correlation provides a useful clue for companies. AI seems to integrate better when not presented as an isolated application, but rather as a tool supported by training, clear objectives, and ongoing assistance.
The paradox of feeling less productive despite using more AI
The most striking result in the report appears when asking about productivity.
16% of those who use AI almost daily strongly agree that they haven’t been as productive as they could be lately. This percentage decreases progressively to 5% among people who have never tried these tools.
This means daily users are over three times, and roughly four times according to ADP’s interpretation, more likely to acknowledge lower-than-desired productivity than those who don’t use AI.
The comparison may seem contradictory. If AI summarizes documents, drafts replies, analyzes data, or automates processes, one might expect its users to feel they work more efficiently and effectively.
ADP notes that it doesn’t have an objective measure of each respondent’s performance. Therefore, it cannot determine if these individuals actually produce less or if technology is changing their perception of their work quality.
A possible explanation is that employees delegate quick, visible tasks to AI to devote more time to complex, strategic, or long-term projects. These tasks often show less immediate results but may hold more organizational value.
Another factor could be inflated expectations. When a tool promises to accelerate writing, analysis, or coding, any improvement that doesn’t meet that promise might feel insufficient.
Review work can also play a role. An AI-generated response might not be ready for immediate use. Verifying data, correcting errors, adjusting tone, reviewing code, or fact-checking sources can consume time that was initially saved.
AI might also introduce new tasks: learning to craft prompts, selecting models, monitoring confidential data, documenting processes, or determining when an automatic reply is sufficiently reliable.
The report does not specify which of these factors weighs the most. It does show that simply adopting tools does not guarantee employees will perceive a clear improvement in their performance.
Spain progresses but remains behind the leading markets
In Spain, 15% of workers report using generative AI almost daily. Another 17% had not tried these tools at the time of the survey.
Spain’s daily usage is below the 20% worldwide average. It also lags behind leading countries: India with 41%, Nigeria with 39%, and Vietnam with 36%.
Brazil reaches 31%, the United Arab Emirates 30%, and Indonesia 29%. Among major Western economies, the US reports 19%, the UK 20%, Germany 14%, and France 11%.
Japan has the lowest daily usage among the included countries, at 8%. It also has the highest percentage of workers who have never tried AI, at 43%. Conversely, only 2% of workers in India and China report never having used it.
Differences may relate to factors such as age demographics, economic structure, access to digital tools, job types, and the pace at which companies have implemented generative solutions.
The report also shows that familiarity reduces some demographic gaps. Men express slightly more optimism about AI’s impact than women, but this difference nearly disappears among frequent users. A similar pattern occurs with age: optimism tends to decrease among older workers, but the gap narrows among daily users of the technology.
Using AI also influences perceptions of job security
The more frequently AI is used, the higher the declared confidence that one’s job will not be eliminated.
This relationship doesn’t necessarily mean habitual users are objectively protected from restructuring. It might indicate they feel more prepared to adapt, work in companies with more technological investment, or hold roles where AI complements their functions.
Knowledge workers, in fact, report the greatest sense of job security in the study. 30% firmly believe their position won’t be eliminated, compared to 18% of those with specialized tasks and 16% of repetitive task workers.
Overall, confidence remains low. Only 22% of global workers believe their employment is secure, with Spain at 24%.
AI use might boost personal confidence, but it doesn’t eliminate worries about automation. Young workers, despite being the most frequent users, show considerable caution regarding AI’s potential impact on employment.
ADP cites studies linking higher AI exposure with reduced job creation for youth in certain sectors. However, the report refrains from making definitive predictions about how many jobs will disappear or be created.
Training matters more than just issuing licenses
The report suggests that organizations should help workers understand where AI adds value and how to incorporate it into their daily activities.
Purchasing licenses and enabling an assistant doesn’t address issues like response quality, confidentiality, access permissions, human review, or performance measurement. It also doesn’t ensure employees know which tasks to delegate and which require professional judgment.
Training should go beyond teaching how to write prompts. It must also cover verifying results, handling personal data, intellectual property, model errors, and the limitations of each tool.
Companies will also need to reassess how they measure productivity. Counting emails sent, completed documents, or lines of code may be less useful when parts of those tasks are automated.
Value might shift toward decision quality, problem-solving speed, customer service, error reduction, or capacity for more complex projects.
Adoption of AI is advancing faster than many organizations can adapt. ADP’s data reflect a workforce increasingly using technology, feeling more engaged, and experiencing less stress, but without a clear perception of productivity improvements as promised.
Frequently Asked Questions
How many workers use AI in their jobs?
Half of respondents use AI several times a week. 20% say they use it almost every day as part of their work routine.
Does AI make employees more productive?
The report does not measure actual productivity. Daily users show more engagement and less stress, but are also more likely to feel they could be more productive.
How prevalent is AI use in Spain?
In Spain, 15% of workers use generative AI almost daily. An additional 17% had not tried these tools at the time of the survey.
Which countries lead in daily AI usage?
India ranks first with 41%, followed by Nigeria with 39%, and Vietnam with 36%.
