![\"\"](\"https:\/\/i0.wp.com\/www.technollama.co.uk\/wp-content\/uploads\/2025\/04\/research.jpeg?resize=750%2C750&ssl=1\")
<\/a><\/p>\nI\u2019m currently on sabbatical finishing a couple of papers, and I wanted to find a figure for a footnote. How many daily users are there for generative AI apps? Simple question, but it turns out that it is quite difficult to find out! I searched through a few sources, but the figures were pretty vague, varying from just 12 million, up to 500 million. So I asked several AI models. I have access to Gemini 2.5, which is quickly becoming the model of choice for many power users, and I really hadn\u2019t tested it yet. So I asked a simple question. \u201cHow many users around the world are using generative AI daily?\u201d<\/p>\n
I was taken aback by the quality of the output, not only by the research, but also by the manner in which it is displayed, and how the reasoning goes. So I\u2019ve decided to reproduce this report here as I found it fascinating, and I know that not everyone has access to Gemini 2.5. It will still be just a footnote on my article.<\/p>\n
Here it is, you be the judge.<\/p>\n
1. Executive Summary<\/b><\/h2>\n
This report assesses the scale of global daily active usage for generative artificial intelligence (GenAI) tools as of early 2025. Pinpointing an exact figure remains challenging due to inconsistent reporting standards, reliance on proxy metrics, and inherent limitations in survey methodologies. However, synthesizing available data from platform disclosures, market research, user surveys, and web analytics provides a basis for estimation.<\/span><\/p>\nThe core finding suggests that the global unique daily active user (DAU) base for generative AI tools likely falls within the <\/span>range of 115 million to 180 million individuals<\/b>. This estimate is heavily influenced by the market dominance of OpenAI\u2019s ChatGPT, which boasts reported weekly active users (WAU) of 400 million.<\/span>1<\/span> Rapid adoption cycles, particularly among younger demographics and specific professional segments (like marketing and IT), alongside increasing integration into existing workflows and platforms (Microsoft 365, Google Workspace), are key drivers of this engagement.<\/span>1<\/span><\/p>\nSignificant measurement challenges persist. Platforms primarily report WAU, Monthly Active Users (MAU), website visits, or subscriber counts, necessitating estimations to derive DAU figures.<\/span>1<\/span> Survey data varies widely depending on the definition of \u201cuse\u201d versus \u201cinteraction,\u201d geographic scope, and methodology.<\/span>1<\/span> Despite these caveats, the overall trend is one of undeniable and rapid growth in frequent GenAI usage, indicating a technology achieving significant daily penetration, albeit with substantial room for further expansion.<\/span><\/p>\n2. The Challenge of Measuring Global Daily Generative AI Usage<\/b><\/h2>\n
Accurately quantifying the number of individuals using generative AI tools on a daily basis globally is fraught with complexity. Several fundamental challenges impede the calculation of a precise DAU figure.<\/span><\/p>\n\n- Defining \u201cActive Use\u201d:<\/b> A primary hurdle is the lack of a universally accepted definition of \u201cactive use\u201d for generative AI. Does interacting with an AI-powered email spam filter count as daily use?<\/span>30<\/span> Does opening a chatbot application suffice, or must a user submit a query? Is usage measured by sustained engagement time or simply any interaction within a 24-hour period? Different data sources implicitly use varying definitions. For instance, surveys reporting that 27% of Americans have <\/span>daily interactions<\/span><\/i> with AI likely encompass passive systems <\/span>26<\/span>, whereas studies focusing on <\/span>frequent daily use<\/span><\/i> of specific tools like ChatGPT report much lower figures (e.g., 7% in the US).<\/span>1<\/span> This ambiguity makes direct comparison across sources problematic.<\/span><\/li>\n
- Lack of Standardized Reporting:<\/b> Major generative AI providers like OpenAI, Google, and Microsoft do not consistently report DAU figures. Instead, they release various metrics such as WAU, MAU, subscriber numbers, platform visits, or app downloads at irregular intervals.<\/span>1<\/span> This fragmented reporting landscape prevents a straightforward aggregation or comparison of daily usage across the leading platforms, forcing reliance on estimation and inference.<\/span><\/li>\n
- Proxy Data Reliance:<\/b> Consequently, analysts must often rely on proxy metrics to estimate DAU. WAU, MAU, website visits, and app downloads serve as indicators of user base size and engagement, but converting these into reliable DAU figures requires assumptions. Standard WAU\/DAU or MAU\/DAU conversion ratios observed in other sectors (like social media, where Facebook\u2019s DAU\/MAU ratio is often cited as 60-70%) may not directly apply to generative AI. GenAI usage might be more task-oriented and sporadic for some users compared to the habitual engagement patterns of social platforms. Applying a fixed conversion ratio (e.g., assuming DAU is 30%, 40%, or 50% of WAU\/MAU) introduces a significant margin of error, making the final DAU estimate highly sensitive to the chosen methodology. This reliance on proxy data means any derived DAU figure for GenAI is inherently less precise than metrics directly reported by platforms for other application types.<\/span><\/li>\n
- Survey Limitations:<\/b> While surveys offer valuable insights into usage frequency, they come with inherent limitations:<\/span><\/li>\n<\/ul>\n
\n- Self-Reporting Bias:<\/span><\/i> Respondents may inaccurately recall or report their usage frequency.<\/span><\/li>\n
- Question Phrasing:<\/span><\/i> Variations in how questions are worded (e.g., \u201cinteract with AI\u201d vs. \u201cuse ChatGPT,\u201d \u201cdaily\u201d vs. \u201cfrequently\u201d) can lead to vastly different results.<\/span><\/li>\n
- Geographic Skew:<\/span><\/i> Many publicly available surveys focus heavily on the United States <\/span>1<\/span> or a limited set of developed countries <\/span>31<\/span>, making global extrapolation difficult.<\/span><\/li>\n
- Representativeness:<\/span><\/i> Online survey panels may over-represent digitally savvy individuals and under-represent older, less connected populations.<\/span>34<\/span><\/li>\n
- Timeliness:<\/span><\/i> The rapid evolution of GenAI means survey data can become outdated quickly. The contrast between Pew Research finding 27% of US adults interact with AI daily\/several times a day <\/span>28<\/span> and Reuters Institute finding only 7% use ChatGPT daily <\/span>1<\/span> exemplifies these challenges.<\/span><\/li>\n<\/ul>\n
\n- Web Analytics Caveats:<\/b> Data from web analytics firms like Similarweb provides valuable insights into website traffic but must be interpreted cautiously.<\/span>35<\/span> These platforms measure <\/span>visits<\/span><\/i> and <\/span>unique visitors<\/span><\/i> to specific domains, which do not directly equate to active users performing generative tasks. A single user can generate multiple visits within a day or month. Furthermore, discrepancies can exist between web analytics figures and user numbers reported by the platforms themselves.<\/span>39<\/span> Analyzing traffic sources (direct, organic search, referral, social) offers clues about user behavior \u2013 for instance, high direct traffic suggests intentional use <\/span>15<\/span> \u2013 but does not provide a precise DAU count.<\/span><\/li>\n<\/ul>\n
3. Survey Snapshots: Gauging Usage Frequency<\/b><\/h2>\n
Surveys provide a direct, albeit imperfect, lens into how often individuals report using generative AI tools. Findings vary significantly based on methodology, population sampled, and the specific questions asked, but collectively they paint a picture of emerging usage patterns.<\/span><\/p>\nUS-Focused Surveys:<\/b><\/p>\n
Several studies have focused on the US market, revealing a spectrum of engagement levels:<\/span><\/p>\n\n- A Pew Research Center report (April 2025, based on 2024 data) found that 27% of U.S. adults claim to interact with AI (broadly defined) \u201calmost constantly\u201d or \u201cseveral times a day\u201d.<\/span>28<\/span> However, the same study noted that only one-third of US adults had ever used an AI chatbot like ChatGPT.<\/span>28<\/span> Interestingly, AI experts surveyed believed a much higher proportion (79%) of the public interacts with AI daily or constantly.<\/span>28<\/span><\/li>\n
- An Elfsight blog citing a McKinsey survey (August 2024) reported that nearly 40% of U.S. adults aged 18-64 had used generative AI to some extent. Among these respondents, about one-third reported using it daily or at least a few times per week. This survey also highlighted that daily use was more common at work (10.6% of respondents) than at home (6.4%).<\/span>4<\/span><\/li>\n
- Data compiled by Photoaid (likely from 2023\/2024) indicated that roughly 27% of Americans report daily interactions with AI.<\/span>26<\/span><\/li>\n
- DemandSage, citing Authority Hacker, offered a more specific figure for workplace usage, stating that 9% of workers in the United States use Generative AI daily.<\/span>27<\/span><\/li>\n
- Salesforce\u2019s ongoing research series found that 45% of the surveyed US population uses generative AI, though specific daily frequency for the entire sample wasn\u2019t provided. It did note a tendency for users to become \u201csuper-users,\u201d implying frequent engagement among adopters.<\/span>5<\/span><\/li>\n
- A FlexOS report, likely targeting professionals, found that 14.5% of their respondents used GenAI daily, with an additional 17% using it multiple times per week, and 8% weekly. This sums to approximately 40% using these tools at least weekly.<\/span>46<\/span><\/li>\n<\/ul>\n
Global\/Multi-Country Surveys:<\/b><\/p>\n
Surveys with broader geographic scope reveal significant international variations and often lower daily usage figures for specific tools compared to general AI interaction rates:<\/span><\/p>\n\n- The Reuters Institute\u2019s 2024 research across six developed countries (US, UK, France, Germany, Japan, Argentina) found that <\/span>frequent daily use<\/span><\/i> of ChatGPT specifically was rare. It was highest in the USA at 7%, followed by Argentina (5%), UK and France (both 2%), and Japan (1%).<\/span>1<\/span> This study emphasized that many who reported using GenAI had only done so once or twice, suggesting it hadn\u2019t yet become a routine part of internet use for the majority.<\/span>31<\/span> Averaging across the six countries, only 5% reported using GenAI to get news.<\/span>31<\/span><\/li>\n
- Salesforce\u2019s research highlighted stark geographic differences in overall usage (not necessarily daily). While 45% of the US sample reported using GenAI, the figure surged to 73% in India but dropped to 49% in Australia and only 29% in the UK.<\/span>5<\/span> This underscores the unreliability of extrapolating global trends solely from US data.<\/span><\/li>\n<\/ul>\n
Demographic Insights from Surveys:<\/b><\/p>\n
Surveys consistently reveal patterns related to age, profession, and education:<\/span><\/p>\n\n- Age:<\/span><\/i> Younger generations, particularly Millennials (born ~1981-1996) and Gen Z (born ~1997-2012), are the primary drivers of adoption and frequent use.<\/span>1<\/span> Salesforce reported 65% of GenAI users are Millennials or Gen Z.<\/span>5<\/span> Pew Research found 43% of US adults aged 18-29 have used ChatGPT, compared to only 6% of those 65 and older.<\/span>1<\/span> Conversely, some studies suggest high weekly use of virtual assistants (a form of AI) among older groups as well.<\/span>10<\/span><\/li>\n
- Profession\/Industry:<\/span><\/i> Adoption and usage frequency are higher in specific sectors like IT, Marketing & Advertising, Customer Support, and Sales.<\/span>3<\/span> A Section survey identified a small segment of \u201cAI Experts\u201d (1% of the workforce in US\/UK\/Canada) where 67% reported using AI daily.<\/span>3<\/span> A Planable survey found around 40% of marketers use AI tools daily, with another 17% using them weekly.<\/span>21<\/span><\/li>\n
- Education:<\/span><\/i> Higher levels of education generally correlate with greater awareness and usage of generative AI tools.<\/span>10<\/span><\/li>\n<\/ul>\n
Consolidated Survey Findings:<\/b><\/p>\n
The diverse findings from surveys underscore the difficulty in arriving at a single, simple percentage for daily GenAI use. However, consolidating key data points reveals important nuances:<\/span><\/p>\n\n\n\n| Survey Source<\/b><\/td>\n | Population Surveyed<\/b><\/td>\n | Metric (% Daily Use\/Interaction)<\/b><\/td>\n | Specific Tool<\/b><\/td>\n | Date\/Report Year<\/b><\/td>\n | Geographic Scope<\/b><\/td>\n | Snippet IDs<\/b><\/td>\n<\/tr>\n\n| Pew Research<\/span><\/td>\n | US Adults<\/span><\/td>\n | 27% (Interaction)<\/span><\/td>\n | AI (General)<\/span><\/td>\n | 2024 data\/2025 Rpt<\/span><\/td>\n | US<\/span><\/td>\n | 28<\/span><\/td>\n<\/tr>\n\n| Reuters Institute<\/span><\/td>\n | 6-Country Online Pop.<\/span><\/td>\n | 7% (US), 2% (UK\/FR), 1% (JP)<\/span><\/td>\n | ChatGPT (Daily Use)<\/span><\/td>\n | 2024 Rpt<\/span><\/td>\n | US, UK, FR, DE, JP, AR<\/span><\/td>\n | 1<\/span><\/td>\n<\/tr>\n\n| Elfsight\/McKinsey<\/span><\/td>\n | US Adults (18-64)<\/span><\/td>\n | ~11% (of respondents)<\/span><\/td>\n | GenAI (Daily\/Weekly)<\/span><\/td>\n | Aug 2024<\/span><\/td>\n | US<\/span><\/td>\n | 4<\/span><\/td>\n<\/tr>\n\n| Photoaid<\/span><\/td>\n | US Adults<\/span><\/td>\n | ~27% (Interaction)<\/span><\/td>\n | AI (General)<\/span><\/td>\n | 2023\/2024 data<\/span><\/td>\n | US<\/span><\/td>\n | 26<\/span><\/td>\n<\/tr>\n\n| DemandSage\/Auth. H.<\/span><\/td>\n | US Workers<\/span><\/td>\n | 9%<\/span><\/td>\n | GenAI (Daily Use)<\/span><\/td>\n | 2024 data<\/span><\/td>\n | US<\/span><\/td>\n | 27<\/span><\/td>\n<\/tr>\n\n| FlexOS<\/span><\/td>\n | Workers (likely professionals)<\/span><\/td>\n | 14.5% (Daily), 17% (Multi-Weekly)<\/span><\/td>\n | GenAI Tools<\/span><\/td>\n | 2024 Rpt<\/span><\/td>\n | Likely Global\/US Focus<\/span><\/td>\n | 46<\/span><\/td>\n<\/tr>\n\n| Section Survey<\/span><\/td>\n | Workforce (US\/UK\/CA) \u2013 \u201cAI Experts\u201d segment<\/span><\/td>\n | 67%<\/span><\/td>\n | AI (Daily Use)<\/span><\/td>\n | 2024 data<\/span><\/td>\n | US, UK, CA<\/span><\/td>\n | 3<\/span><\/td>\n<\/tr>\n\n| Planable<\/span><\/td>\n | Marketers<\/span><\/td>\n | ~40% (Daily), 17% (Weekly)<\/span><\/td>\n | AI Tools<\/span><\/td>\n | 2024 Rpt<\/span><\/td>\n | Likely US\/Global<\/span><\/td>\n | 21<\/span><\/td>\n<\/tr>\n\n| AIPRM<\/span><\/td>\n | US Adults<\/span><\/td>\n | 29.5% (Email Spam Filters Daily)<\/span><\/td>\n | Specific AI Types<\/span><\/td>\n | 2024 Survey<\/span><\/td>\n | US<\/span><\/td>\n | 30<\/span><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n (Table 3.1: Summary of Key Survey Findings on Generative AI Usage Frequency)<\/span><\/i><\/p>\nThe data presented in Table 3.1 reveals a significant divergence between figures reporting broad \u201cAI interaction\u201d and those measuring specific \u201cGenerative AI tool daily usage,\u201d particularly for chatbots. The former often yields higher percentages (like 27%), potentially inflated by passive or embedded AI systems such as email spam filters.<\/span>30<\/span> In contrast, surveys asking directly about daily use of tools like ChatGPT show markedly lower figures, often in the single digits to low double digits, especially outside the US.<\/span>1<\/span> This distinction is critical: while AI touches many lives daily in some form, active, intentional daily engagement with <\/span>generative<\/span><\/i> tools appears less widespread across the general population than interaction figures might suggest.<\/span><\/p>\nHowever, this picture of modest overall daily penetration contrasts with reports indicating that those who <\/span>do<\/span><\/i> adopt these tools tend to use them frequently. The concept of \u201csuper-users\u201d <\/span>5<\/span> and the high daily usage reported among specific segments like \u201cAI Experts\u201d <\/span>3<\/span> or marketers <\/span>21<\/span> suggest a pattern of deep adoption within certain groups. This implies that while universal daily use is far from reality, a dedicated and growing user base engages with these tools regularly, driving the large weekly and monthly active user numbers reported by platforms. The growth appears concentrated within these adopter segments for now.<\/span><\/p>\nFurthermore, the substantial geographic disparities highlighted by surveys <\/span>2<\/span> cannot be overstated. High reported usage in countries like India <\/span>5<\/span> stands in stark contrast to lower figures in the UK <\/span>5<\/span> or Japan.<\/span>31<\/span> This variability, likely driven by factors including digital infrastructure maturity, language model availability and quality, cultural acceptance of AI <\/span>6<\/span>, and regulatory environments, makes any single global DAU average potentially misleading. Extrapolating global usage patterns solely from US-centric data is therefore highly unreliable.<\/span><\/p>\n4. Platform Dominance & User Base Proxies<\/b><\/h2>\nTo estimate global DAU, analyzing the user metrics reported by the dominant generative AI platforms is essential. ChatGPT, Google Gemini, and Microsoft Copilot represent the largest pools of potential daily users, and their reported figures, while varied, serve as crucial proxies.<\/span><\/p>\nChatGPT (OpenAI):<\/b><\/p>\n OpenAI\u2019s ChatGPT has consistently maintained a significant lead in user adoption and engagement since its launch.<\/span><\/p>\n\n- Reported Metrics:<\/b> The most frequently cited and recent metric is <\/span>400 million Weekly Active Users (WAU)<\/b> as of February 2025.<\/span>1<\/span> This represents remarkable growth, doubling from 200 million WAU in August 2024 and quadrupling from 100 million WAU in November 2023.<\/span>1<\/span> Some sources directly report figures around <\/span>122.58 million daily users<\/b> 2<\/span> and over <\/span>1 billion daily queries<\/b> processed.<\/span>2<\/span>
| | | | | | | | | |
![\"\"](\"https:\/\/i0.wp.com\/www.technollama.co.uk\/wp-content\/uploads\/2025\/04\/research.jpeg?ssl=1\")
<\/a><\/p>\n