The clearest signal in Meltwater and We Are Social’s Digital 2027 report is a widening split: generative AI has reached billions of monthly users as traffic to major websites has fallen. The 900-plus-page report, released this October, tracks global digital habits, with additional topic reports by author Simon Kemp. Its figures describe a digital audience adopting new tools while some established measures of web use move in the other direction.
That tension deserves attention, but it doesn’t settle the question of what’s causing the change. The report presents AI use alongside lower traffic to major domains and points to mobile apps as another possible factor. Those patterns can coexist without showing that one directly caused the other.
AI use is broad, but paid subscriptions are rare
An estimated 3.24 billion people, or 38.9% of the world’s 8.32 billion population, use generative AI tools each month. That represents roughly 1.2 billion more users, though the report cautions that people may hold multiple accounts, including separate ones for personal and work use. The estimate therefore points to the scale of monthly activity, but it shouldn’t be read as a precise count of distinct people who have tried a tool or use one regularly.
In the Philippines, 94.4% of internet users aged 16 and older said they used AI tools monthly. Only 5.5% paid for an AI subscription. Among active AI users in that age group, 71.17% use the tools to find information, 37.6% seek help with problems, and 32.8% look for personalized recommendations. The gap between reported use and paid subscriptions suggests that broad exposure hasn’t translated into widespread payment, at least in these figures. It also leaves open what respondents mean by “use”: the reported tasks range from information seeking to recommendations, and don’t by themselves show how often or how deeply people rely on AI.
Public confidence is another part of the picture. Just 27.5% of Philippine internet users aged 16 and older believe AI benefits society overall. That’s a much narrower measure than monthly use: people can try a tool for practical reasons without deciding that its wider effects are positive. This is more significant than it looks. Adoption alone is a poor proxy for trust.
Philippine digital figures show mixed movement
The country’s population reached 118 million, up 936,000 from the previous year, according to the report’s Philippine figures. There were 149 million cellular connections, equivalent to 126% of the population—a sign that some people have more than one connection. Connections and people aren’t interchangeable measures: a total above the population indicates that the count includes multiple connections for at least some users, rather than showing that every person has a mobile connection.
Internet use stood at 95.9 million people, or 81.3% of the population. The report counted 94.7 million social media user identities, down about 1.1 million. As with global social media totals, identities may include separate personal and work accounts, so they don’t necessarily represent distinct individuals. That caveat matters when comparing social media figures with population or internet-use totals: an account count can track the reach of services, but it can’t automatically tell you how many people are active or how their behavior has changed.
Here’s the thing: these figures point in different directions. Population and cellular connections rose, while reported social media identities fell. Those movements don’t cancel each other out, and the available numbers don’t explain why they happened. They do make a simple story of steady growth harder to defend.
Mobile apps and AI coincide with declining web traffic
Across the world’s 10,000 most visited web domains, monthly visits fell by almost 50 billion between the fourth quarter of 2022 and the second quarter of 2026. Total monthly traffic was down more than 10% since ChatGPT launched in 2022. Time spent on those sites also declined 11.2% over a two-year period, a reduction of more than 6.6 billion hours per month. These are large declines across a defined group of major domains, but they don’t describe every website or every kind of online activity.
The report points to AI use as one possible factor, alongside a shift toward native mobile apps for online tasks. That’s a plausible explanation for some changes in how people reach information and services, but the figures presented here don’t isolate the contribution of either factor. If you’re working in this space, the distinction matters: a drop in visits to a set of websites isn’t proof that people have stopped doing the tasks those sites once served.
And yet the web remains a major channel: visitors loaded just under 3 trillion pages per month across those domains in the second quarter of 2026, more than 2% above the same quarter in 2024. The comparison complicates the decline story. Traffic can fall from an earlier level while remaining substantial, and visits, pages loaded, and time spent measure different kinds of activity. The web’s reach hasn’t vanished; the report describes a change in use that needs more than one metric to understand.
Mobile reach and social identities keep expanding
Worldwide social media user identities reached 5.81 billion, equal to 69.9% of the global population, with the same caveat about duplicate identities. GSMA Intelligence and Ericsson data put the number of unique mobile phone users at 5.86 billion. Ericsson also estimated 7.68 billion smartphones in use, nearly 90% of active mobile handsets and close to 86% of cellular connections; people may own multiple smartphones.
The figures distinguish devices, connections, users, and identities—categories that can overlap but aren’t equivalent. A person may have more than one connection or smartphone, while social platforms may count separate accounts. So the totals show the scale of mobile access and social activity, but they shouldn’t be added together as if each represented a separate person. That said, the broad reach of phones helps explain why app use belongs in any account of changing online behavior.
What the numbers suggest about digital habits
The figures point to a digital audience that’s increasingly comfortable using AI and mobile apps, even as the web’s largest domains receive less traffic. The report offers a useful snapshot of that overlap, but its account of causes is more tentative than its counts: AI use is one possible factor in declining web activity, and a move to native apps is another. The data shared here don’t establish how much each contributes.
For readers, the practical takeaway is to treat reach as only one part of the story. Monthly AI use doesn’t tell you whether someone pays for a service, trusts it, or uses it for consequential decisions. Likewise, fewer visits to major domains don’t show whether people have abandoned the web or shifted tasks elsewhere. The pattern is clear. The explanation is still developing.
The full Digital 2027 report and Simon Kemp’s topic reports are available online.
