Artificial intelligence has already reached more than half of Spanish internet users between the ages of 16 and 75, but the leap from trying out a chatbot to truly weaving it into work and daily life is still pending. IAB Spain’s first Annual Artificial Intelligence Study, produced together with Elogia, puts the share of internet users who have used AI tools over the past year at 52%. The more revealing figure comes next: 85% of those users keep their use basic, occasional or supportive, and only 15% have integrated AI into their routines, automated processes or delegated entire tasks.
AI in Spain: the key figures in 30 seconds
- 52% of internet users aged 16 to 75 have used AI in the past year, with a pronounced generational gap.
- ChatGPT leads with 70% current use and 52% preference, while Gemini reaches 43% usage.
- 85% use AI in a basic way, and 59% don’t use any advanced features.
- 87% stick to free versions, and only 13% pay for AI, either directly or through their employer.
- AI already plays a role in product searches, but 65% wouldn’t let an agent make a purchase on its own.
The IAB Spain study offers a particularly useful snapshot because it doesn’t just measure how many people know about or use artificial intelligence tools. It also looks at which platforms they use, how often, which features they take advantage of, how much they pay, how they use AI in their purchases, how much they trust its answers, and what effects they perceive at work.
The research combines three methodological sources. The main one is an online survey of 1,081 Spanish internet users aged 16 to 75, conducted between May and June 2026. The study also incorporates digital research into searches for AI tools between 2023 and 2026, plus 13 qualitative interviews with professional AI users, carried out through an artificial intelligence agent. The latter are there to provide context, but the report itself warns that their results are indicative rather than representative.
The result paints a picture of a market that has moved past its initial discovery phase. AI already has a broad, frequent presence, but it’s still used mainly as a tool to consult. The real shift will come when a larger share of users move from asking questions to delegating, automating and integrating.
More Than Half of Internet Users Have Already Used AI
52% of Spanish internet users aged 16 to 75 have used some artificial intelligence tool over the past year. IAB Spain compares this figure with earlier data from Spain’s National Observatory for Technology and Society (ONTSI): the agency put the adult population that had ever used AI at 42.1% in 2023, and at more than 50% for those who had used generative AI in 2024.
The 2026 figure refers specifically to the internet-using population aged 16 to 75, so it shouldn’t be read as an identical measurement to the earlier ones. Even so, it shows just how far AI tools have penetrated everyday internet use in Spain.
Age is the main differentiating factor. Generation Z, defined in the study as people aged 16 to 30, reaches 79% penetration. Among millennials, aged 31 to 45, it drops to 64%. Generation X, aged 46 to 60, sits at 47%, while boomers, aged 61 to 75, reach 33%.
| Group | AI use in the past year |
|---|---|
| Generation Z, ages 16-30 | 79% |
| Millennials, ages 31-45 | 64% |
| Generation X, ages 46-60 | 47% |
| Boomers, ages 61-75 | 33% |
| All internet users, ages 16-75 | 52% |
The gap between men and women is much smaller: 51% of men and 53% of women use AI. The average user profile is 46 years old, 58% have a university degree, and 69% are currently employed.
Among those who don’t yet use AI, the average age rises to 54. 36% have a university degree and 60% are employed. 34% of this group say they will probably or definitely start using AI in the next 12 months.
The gap, then, doesn’t disappear just because new users keep arriving. Age still shapes both access to the technology and how intensively it’s used.
ChatGPT Keeps Its Lead, But Gemini Is Growing the Fastest
The Spanish market is highly concentrated. ChatGPT stands out as the dominant platform in awareness, use and preference.
97% of AI users are aware of ChatGPT, 88% say they’ve used it at some point, and 70% currently use it. On top of that, 52% name it as their preferred tool.
Gemini comes in second, with 85% awareness, 60% of users who have tried it, 43% current use and 24% preference.
Microsoft Copilot rounds out the podium in current use, at 20%, although its preference sits at 8%. Its awareness reaches 57%.
The full comparison shows just how far ahead these three platforms are from the rest.
| Tool | Awareness | Has ever used it | Current use | Preference |
|---|---|---|---|---|
| ChatGPT | 97% | 88% | 70% | 52% |
| Gemini | 85% | 60% | 43% | 24% |
| Microsoft Copilot | 57% | 35% | 20% | 8% |
| Meta AI | 56% | 18% | 8% | 3% |
| Claude | 30% | 14% | 7% | 5% |
| Grok | 37% | 15% | 6% | 2% |
| Perplexity | 23% | 10% | 4% | 1% |
| DeepSeek | 23% | 8% | 2% | 1% |
These percentages come from different questions, so they shouldn’t be added together. The table is useful for tracking the journey from awareness to trial, current use and preference.
The evolution of unique users measured by GfK DAM adds another dimension. Between June 2025 and June 2026, the set of AI tools analyzed went from 17.3 to 25.8 million unique users among Spain’s internet-using population, a 49% increase.
ChatGPT grew from 14.33 to 20.42 million unique users, up 42.5%. Gemini was the fastest-growing platform among the three biggest, going from 3.26 to 10.24 million, up 214.4%. Copilot went from 2.22 to 2.78 million, up 25.1%.
Claude’s growth also stands out: from 0.18 to 2.09 million unique users, up 1,090.4%. Meta AI went from 0.22 to 2.02 million, up 834.6%. These are very high growth percentages because they start from small bases.
DeepSeek is the exception: it dropped from 1.00 to 0.75 million unique users, down 24.4%.
The Difference Between Using AI and Knowing How to Use It
The most interesting part of the report for the tech sector comes when IAB Spain stops simply asking whether someone uses AI and starts measuring how they use it.
43% of users say they use it several times a day, and another 11% once a day. 34% use it several times a week and 9% once a week. In total, 86% use AI daily or weekly.
But that frequency doesn’t mean automation.
54% use AI for specific tasks and 31% regularly use it as a support tool. Only 11% say they have integrated AI into their routines. That share drops to 3% when asked about automated processes, and to 2% when it comes to delegating entire tasks to automated systems.
The result is a technological paradox: AI has become a frequent tool without yet becoming a personal work infrastructure for most users.
Page 23 of the study sums up that difference with a split between basic and advanced use. 85% fall into the first category and only 15% into the second. Among heavy users, advanced use reaches 25%, pointing to a direct relationship between frequency and maturity.
The use of advanced features is still limited:
| Feature | Users who use it |
|---|---|
| Delegating intermediate steps | 16% |
| Structured prompts | 15% |
| GPTs or agents | 13% |
| Integrations | 11% |
| Automation via apps or APIs | 7% |
| Building repeatable workflows | 7% |
| No advanced feature | 59% |
That last figure is especially relevant for companies selling AI tools. The challenge is no longer just convincing users to try a model. They also need to help users find use cases that are repeatable and useful enough to fold into their own processes.
The report identifies lack of training as the main obstacle, cited by 31%. Price comes right after, at 30%, followed by inconsistent quality at 28%. Lack of trust reaches 25%, and legal or privacy risks 21%.
The technology is available, but a significant share of users still don’t know how to turn it into a workflow.
Freemium Dominates, and Paying Comes Once AI Becomes a Habit
Mass penetration hasn’t yet translated into an equivalent subscription market either.
87% of users mainly use the free version of these tools. Only 13% pay in some way: 7% through their employer and 6% out of their own pocket.
The difference by intensity of use is clear. Among users who use AI daily, the share who pay rises to 21%.
The study identifies four groups:
| Situation | Percentage |
|---|---|
| Already pays directly | 6% |
| Doesn’t pay but would be willing to | 14% |
| Doesn’t know yet if they’ll pay | 33% |
| Doesn’t plan to pay | 47% |
Add up those willing to pay and those still undecided, and you get 47%. That doesn’t mean there’s a guaranteed 47% of future customers, but rather a group the study considers potentially convertible.
Price also has a fairly low ceiling. Among those who would be willing to pay, two out of three wouldn’t go above €10 a month, and the average price cited by the study is €9.4.
The consequence for the market is clear: growing the number of users doesn’t guarantee growing revenue. The move to paid plans depends on the tool solving problems that are frequent and relevant enough.
AI Is Already Changing Product Searches
Commerce is another area where the study detects a shift in behavior.
45% of AI users have already used it to help with a purchase. They mainly use it to compare products (66%) and to research (57%).
There’s one figure especially relevant for search engines, specialized media and e-commerce: out of every ten product searches these users carry out, 24% go through AI, 64% through search engines, and 12% through other channels.
Among Generation Z and heavy users, the share of product searches done through AI reaches 32%.
That doesn’t mean Google has stopped being the main channel. Search engines still hold 64% of the searches analyzed. What the study shows is that AI already occupies a significant part of a stage that was traditionally dominated by search engines and results pages.
The difference shows up when AI tries to move from recommending to buying.
The Shopping Agent Still Needs Permission
Among people who have used AI during a purchase, 65% wouldn’t let an agent buy on their behalf. 21% would accept that only under supervision, and just 14% would allow a purchase with no conditions attached.
Even among those willing to delegate, financial autonomy is limited.
| Amount they would allow to be spent without review | Percentage |
|---|---|
| €0, would always review | 49% |
| Less than €50 | 28% |
| Between €50 and €200 | 14% |
| Between €200 and €500 | 6% |
| More than €500 | 3% |
The pattern also shifts by category. Travel, technology and home are the areas with the greatest willingness to delegate, at 39%, 37% and 36% respectively. At the other end are children’s products at 7%, eyewear at 11%, and pet products at 13%.
The logic the study reflects is fairly straightforward: the simpler and less sensitive the purchase, the greater the willingness to leave part of the process in AI’s hands.
For e-commerce, that puts the agent in a middle position. AI already influences research and comparison, but it still doesn’t have permission to automatically close most purchases.
95% Double-Check AI’s Answers
Trust is possibly the main limit on AI moving into tasks with greater responsibility.
68% of users believe artificial intelligence will become essential in the future. However, only 29% say they trust it. Between those two figures sits a 39 percentage-point gap.
That caution also shows up in actual behavior. 95% check, to some degree, the information they get from AI. 22% always do so and 32% frequently.
When checking a response, users mainly turn to search engines, used by 68%, and official websites, at 61%. 30% ask someone they trust, and 21% check other users’ reviews.
Only 8% use another AI to double-check a response.
This is especially relevant for the future of search. AI may be displacing some initial queries, but external sources still function as a validation mechanism.
Spanish users don’t seem to be replacing one source with another. They’re adding a new layer and checking its results.
AI Boosts Productivity, But It Can Also Increase Workload
The professional sphere offers some of the study’s clearest results. 48% of users use AI at work, either for professional tasks or combined with personal uses.
Among professionals who use AI, 90% perceive some positive impact. 53% say they work faster, 40% say the quality of their work has improved, and another 40% say it frees up time. On top of that, 21% say they can now do tasks they couldn’t do before.
Generation Z stands out especially on speed: 63% say they work faster thanks to AI.
That optimism lines up with an earlier survey finding that most Spaniards trust AI to help them at work rather than replace them, even if the new IAB Spain data shows that trust hasn’t yet translated into deep day-to-day integration.
The study, however, includes a figure that keeps productivity from being reduced to a simple matter of saving time.
Among those who free up time thanks to AI, 47% spend it on their personal life, while 46% use it to take on more work. 37% put it toward strategic tasks and 26% toward training.
| Where the freed-up time goes | Percentage |
|---|---|
| Personal life | 47% |
| More workload | 46% |
| Strategic tasks | 37% |
| Training | 26% |
These percentages correspond to multiple-choice answers, so they shouldn’t add up to 100%.
The so-called productivity paradox is thus fairly well defined: saving time with AI can improve personal balance, but it can also raise expectations about how much work a person should produce in that same period.
What All This Means for the AI Market
IAB Spain’s data paints a picture of an industry moving at three different speeds.
The first is adoption. There, AI has already won. 52% of Spanish internet users aged 16 to 75 have used it over the past year, and 86% of current users use it at least weekly.
The second is integration. Here, the market is much less developed. Only 15% have gone as far as integrating AI into their routines, automating processes or delegating entire tasks. 59% don’t use any advanced feature.
The third is trust. 68% believe AI will become essential in the future, but only 29% currently trust it.
These three figures help make better sense of where the market stands:
| Indicator | Status in Spain, 2026 |
|---|---|
| Internet users who have used AI in the past year | 52% |
| AI users who use it daily or weekly | 86% |
| Users with basic use | 85% |
| Users with advanced use | 15% |
| Users who don’t use advanced features | 59% |
| Users who pay for AI | 13% |
| Users who believe it will become essential | 68% |
| Users who trust AI | 29% |
| Users who double-check its answers | 95% |
The picture is less dramatic than some forecasts of an immediate wave of autonomous agents, but it’s more useful for understanding where users actually stand.
AI is already woven into the digital habits of millions of people. What’s still unresolved is how to move from a conversation to a process, from a query to automation, and from a free tool to a service users consider reasonable to pay for.
The question of trust also remains unresolved. A user can use ChatGPT or Gemini several times a day and, at the same time, check almost all of its answers before making a decision. That combination of practical reliance and distrust explains a good part of today’s market.
For developers and tech companies, this is probably one of the study’s most relevant findings. AI’s next phase of growth doesn’t just depend on gaining more users. It depends on getting those users to do more with it.
Frequently asked questions
How many Spanish internet users use artificial intelligence?
52% of Spanish internet users aged 16 to 75 have used AI tools over the past year, according to the study by IAB Spain and Elogia.
Which AI is used the most in Spain?
ChatGPT leads with 70% current use among AI users and 52% preference. Gemini comes in second with 43% current use and 24% preference.
How many people pay for AI tools?
13% of users pay for AI in some way: 7% through their employer and 6% directly. Among users who use AI daily, that share rises to 21%.
Do Spanish users trust AI’s answers?
Trust is still limited. 29% say they trust AI, while 95% check, to some degree, the information they get from these tools.
