> **来源：[研报客](https://pc.yanbaoke.cn)** # Our life with Al: From experimental to essential Third annual global report from Google and Ipsos January 2026 # Foreword. In 2025, people moved past experimentation and embraced AI in their everyday lives. This shift from exploring with AI to regularly using it as a helpful tool is profound: people now use AI not only to explore what it can do, but to support them in their own lives, both personally and professionally. For the third year in a row, Google partnered with Ipsos—conducting 21,000 interviews across 21 countries—to continue to understand public attitudes toward AI. For the first time, our data shows that the world has decisively crossed the adoption threshold—with majorities of the population across all but a few countries reporting using an AI chatbot. AI is no longer a futuristic curiosity; it is an essential tool relied upon by people worldwide. In 2023, people navigated curiosity and uncertainty around AI following the emergence of Large Language Models (LLMs) and chatbots. In 2024, AI unlocked new benefits for society, from improving flood forecasting to finishing sentences. In 2025, people embraced AI in their everyday lives and experienced the benefits of AI firsthand. Now: - Interest in learning more about AI has grown Using AI to support learning has emerged as a new top AI use case People are using AI to plan, make decisions, and advance their careers - More people believe society, workers, and people like them will benefit from AI As with any new technology, people have concerns about AI—including real worries about AI's impact on workers and concerns about AI in the hands of bad actors. More pessimistic attitudes are pronounced in the West (particularly in the United States and Canada and in some European countries), but globally, excitement about AI continues to outweigh concerns about potential risks. The global public continues to look to governments to safely maximize AI's benefits for individuals and society. Trust in both governments and technology companies to responsibly develop AI is high—and consistent year over year. As we look ahead to another exciting year with AI, we hope this report can continue to contribute to conversations about AI's value for society and individuals and the importance of working together to safely advance AI innovation. # Contents. 1 Helpfulness in the hands of more people. More people are using AI and the technology is penetrating into people's lives more than ever before. 2 Lifelong learning powered by AI. Teachers, students, and parents are super users of AI and the most bullish on AI's benefits for education. 3 Experiencing is (still) believing. Countries and people who use AI more are the most optimistic about AI's potential. 4 AI's impact. While perceptions of how AI will impact people have improved, the public has concerns about AI's impact on workers. 5 Innovating responsibly—together. The public trusts governments and AI companies and expects them to work together to serve citizens. In 2025, AI crossed the adoption threshold, with majorities of the global public reporting having used an AI tool or chatbot. # Helpfulness in the hands of more people. # Practical power. In 2024, most people looked to AI for entertainment or exploration. In 2025, people started using AI more as a practical tool. AI supported learning and deep understanding (74% of AI users overall), saving time (65%), and even finding the right words to navigate a tough situation (64%). More than half (52%) of users have now used AI to assist in their daily lives, and a not insignificant 4 in 10 have used AI to explore a new business venture or career change (42%) or to help navigate a major life event or decision (40%). AI is now a helpful and meaningful part of life—whether people use it for meal planning or generating the perfect invitation for a child's birthday party. How AI applications were used in the past 12 months # Adoption ascending. Usage of and interest in AI has increased since we began tracking in 2023. Two thirds of the global public (66%) now says they have used an AI tool. Since 2023, self-reported chatbot usage has jumped 24 points—from $38\%$ to $62\%$ of the global public.* About 1 in 5 adults (21%) say they use AI "a lot" in their work or life. Outside of this core group of very regular users, there is real appetite to learn and do more: over three-quarters of the global public (77%) are interested in learning more about AI, up from 68% in 2023. AI adoption # Who has adopted AI? AI adopters (62% overall have used a chatbot in the last year) are more likely to be under 35 (79%), higher educated (75%), parents (74%), and have higher household incomes (68%). Students $18+$ (79%) and teachers (80%) use AI more than the public overall. Those who haven't used chatbots (33% overall) tend to be Baby Boomers (58% have never used an AI chatbot) and people who are not currently part of the workforce (45%). Relative to the general public, non-users are also slightly more likely to have lower educational attainment (37%) and lower household income (also 37%). Notably, the gender gap in AI usage appears to be smoothing out: $64\%$ of men have used AI chatbots in the last year, compared to $61\%$ of women (there was a 10-point gender gap in usage in both 2023 and 2024). AI use by demographic # Lifelong learning powered by AI. # The new super users. Students, teachers, and parents see real value in AI. They use AI at higher rates than the overall public and have emerged as 2025's AI "super users." Students: $85 \%$ of students $18+$ are using AI tools. Students use it to help with schoolwork ( $83 \%$ of student users), understand complex topics ( $78 \%$ ), manage day- to- day life tasks like trips, meals, or workouts ( $54 \%$ ), and make decisions ( $42 \%$ ). Teachers: $81 \%$ of teachers report using AI tools, far surpassing the global average (66% of the global public reports using an AI tool). Some of educators' top uses for AI are learning something new or understanding a complex topic (77% of teacher users) and saving time (75%). Parents: $76 \%$ of parents say they use AI tools, especially to learn something new (77% of parent users) or to assist with work (73%). Nearly half (49%) of parents report using AI to explore changing careers, making more money, or starting a new business. Most importantly, rather than fearing cognitive decline, teachers, students, and parents believe AI is having a positive impact on how we learn. # Perceived positive impact on the way we learn # The public sees AI as a net positive for education. AI is seen as a value-add for individual learners—two thirds of the global public feel AI is having a positive impact on the way we access information (66% positive) and on the way we learn (65%). Even in countries like the United States, where attitudes about AI are the most negative, people say AI is having a positive impact on access to information (55% positive / 22% negative) and learning (45% positive / 30% negative). Using AI for education is also seen as a net benefit for society: the public believes primary and secondary students (70% likely to benefit globally), university students (79%), and educators (75%) are likely to benefit from AI. This is true in the United States, Canada, and Europe and even more pronounced in the Asia-Pacific region and in emerging markets. Despite these benefits for individuals and society, many have concerns that AI could erode critical thinking and reduce teaching quality. Teachers, however, believe AI will improve student outcomes (63% improve / 37% worsen) and improve teaching quality (67% improve / 33% worsen). # Likelihood of benefitting from AI Educators University students Primary and secondary students # Inequality and excellence in education. Outside of the United States, Canada, and Europe, the global public echoes student and teachers' positive attitudes about AI for education. In emerging markets, people believe AI is more likely to improve student outcomes by supporting personalized learning (63%) than worsen student outcomes by eroding critical thinking (37%). Overall, countries identified as having greater inequality are much more likely to think AI will improve teaching quality and student outcomes than countries with lower inequality.* Countries that excel in education are also eager to tap into AI. In South Korea, Japan, and Singapore, the only countries we surveyed that boast $500+$ PISA scores in science, reading, and math, attitudes about AI's role in education are similarly positive.\*\* Perceived impact of AI on education Base: all respondents. "Q. Do you think the use of AI in education will..." *Gini inequality coefficients from World Bank. Low inequality: India, United Arab Emirates, Belgium, Poland, Ireland, Canada, France; moderate inequality: Australia, Italy, Spain, South Korea, Germany, United Kingdom, Japan; high inequality: South Africa, Brazil, Singapore, Mexico, Argentina, United States, Nigeria. **PISA scores from World Population Review. # Experiencing is (still) believing. # Usage continues to drive excitement about AI. In our first global report from 2023, we found that experiencing is believing when it comes to AI. When people use AI, excitement about what the technology can do outweighs and assuages concerns about potential risks. This continues to be the case. Among those who have used AI, $69\%$ are more excited about the possibilities of AI than concerned about the risks. Among those who say they use AI "a lot," more than 8 in 10 (86%) are excited. Knowing more about AI is also correlated with lower concern: among those who say they know "a lot" about AI, $80 \%$ are excited and just $20 \%$ are concerned. # Excitement vs. concern about AI # Excitement mirrors usage globally. As we have seen in previous years, excitement about AI varies by region—people in the West are less excited, while people in Nigeria, the United Arab Emirates, and India are particularly excited and optimistic. There is a clear correlation between self-reported AI usage and excitement about the technology: countries that have higher self-reported chatbot usage are more likely to report being excited about AI than countries with lower usage, like the U.S., Canada, Japan, and France. # AI use in the past 12 months and excitement (ordered by excitement) Base: all respondents. "Q. In the last twelve months, have you used generative AI tools like ChatGPT, Gemini, or Claude, or used AI to generate images, video, or audio?" Base: all respondents. "Q. When you think about Al, are you more...?" [Excited about the possibilities shown] # AI's impact. # Greater positive impact. In the last three years, public perceptions of AI's impact on people have improved, with more of the public believing workers, society, and people like them are likely to benefit from AI. More people also think disabled people and underrepresented groups are likely to benefit from AI. Positive perceptions of AI's societal impact are highest in emerging markets (83% believe society as a whole is likely to benefit) and the Asia-Pacific region (76%), followed by Europe (65%) and North America (57%). Likelihood of benefitting from AI # Weighing Al's impact on workers. Globally, two thirds (66%) of the public think workers are likely to benefit from AI, a 13-point increase from 2023. Perceptions are most positive in emerging markets (78% likely to benefit), but even in the United States and Canada, where attitudes have been consistently more pessimistic about AI, 53% believe workers are likely to benefit from AI, compared to 32% who believe they are not. While the public feel AI will have a positive impact on workers overall, they also have very real concerns about how AI will impact jobs. Those who feel AI changing jobs and industries is a "good thing" still outpace those who say it's a bad thing, but the figure has declined since last year (49% good thing / 32% bad thing, compared to 58%/24% last year). Currently, the public is split 50/50 on whether AI's benefits outweigh potential job disruption. However, among those who have used AI for work, and among younger workers, more feel that AI will ultimately help workers than hurt workers. Despite concerns, people continue to believe fostering advances in science, medicine, and other fields through AI innovation is more important (58% globally, compared to 59% last year) than protecting impacted industries through AI regulation (41%, level with last year). The push towards innovation holds true even in countries like the U.S. that tend to be more pessimistic (53% innovation / 44% regulation). # AI's impact on the workplace AI in the workplace will eliminate jobs and disrupt industries, ultimately hurting workers AI in the workplace will create new jobs and new ways of working, ultimately helping workers # Balancing benefits and potential downsides from Al. The public recognizes that—as with most new technologies—there are benefits and downsides to AI. Take online content, for example: AI enables faster threat detection and response (which $38\%$ of the public feel is a very promising application), but bad actors have also used AI to create deepfakes and spread disinformation ( $53\%$ very concerning). The same goes for AI's impact on workers and the workforce, as we note earlier—yes, AI can make workers more effective and allow them to focus on more rewarding work ( $37\%$ very promising), but the public is concerned that AI could also lead to job losses ( $48\%$ very concerning). While some issues will be more complex to navigate, issues like accessibility and science continue to be identified as particularly promising—and lower risk (accessibility: $49\%$ very promising/ $42\%$ very concerning; science: $49\%$ very promising/ $39\%$ very concerning). Very promising applications of AI (8-10 out of 10) # Innovating responsibly—together. # Trust in AI companies to introduce AI responsibly remains stable year over year. A majority of the global public continues to think that AI companies are introducing AI responsibly (57%), and three-quarters (74%) say they have confidence in tech companies to oversee AI in the best interest of the public. # Trust and confidence in technology companies Base: all respondents. "Q. How much confidence do you have, if any, in the following to oversee the development of AI in the best interest of the public?" [Technology companies; total confidence shown] Base: all respondents. "Q. Which of the following do you agree with more, even if neither is exactly right?" [Technology companies are introducing AI responsibly, acknowledging both the potential benefits and potential downsides A Technology companies are introducing AI irresponsibly and not considering the consequences]" # Governments are trusted, but could be doing more. Overall, trust in governments to oversee the development of AI in the best interest of the public is high (57% confident overall, down slightly from 61% 2024 but in line with 2023). Across regions, majorities of the public still believe governments are doing a good job capturing the benefits of AI (63% globally, in line with last year). Notably, however, 4 in 10 Europeans (40%) still believe their country is putting limitations on AI that will prevent their economy from benefiting from the technology (down only slightly from $41 \%$ last year). Citizens in the West are less inclined to feel their government is doing a good job using AI to streamline government operations and better serve citizens, and up to a third do not know whether their government is doing a good job using AI. My country's government is effectively leveraging AI Base: all respondents. "Q. Please indicate how much you agree or disagree with the following statement: "I think my country's government is doing a good job using AI to streamline government operations and better serve citizens." # Working together to unlock AI's benefits. As in previous years, the public still expects governments and tech companies to work together to not only oversee AI development, but to put AI to work for government and citizens. Majorities of the public—including in the United States and Europe—want to see governments using AI to help citizens access public services, unlock insights from complex data, and make systems more secure from cyber attacks. # Support for AI investment and government use Support governments using AI to help citizens access public services Support governments using AI to unlock insights from complex data Support governments using AI to make their systems more secure from cyber attacks # Methodology. These are findings from a survey conducted by Ipsos between September 22 and October 10, 2025, on behalf of Google. For this survey, a sample of roughly 1,000 adults $18+$ , representative of the general population of each in Australia, Belgium, France, Germany, Ireland, Italy, Japan, Poland, Singapore, Spain, South Korea, and the United Kingdom were interviewed online using Ipsos online panels. Survey samples of roughly 1,000 adults $18+$ representative of the online population of each in Brazil, India, Mexico, Nigeria, South Africa, Argentina, and United Arab Emirates were also interviewed using Ipsos online panels. A sample of 1,000 adults age $18+$ who are residents of the United States were interviewed online via the KnowledgePanel®. The survey has a credibility interval of plus or minus 3.8 percentage points for all respondents in each country. Trend data is from the 2023 and 2024 Google / Ipsos Our Life with AI studies. The 2023 Google / Ipsos Our Life with AI study was fielded online between October 19 and November 6, 2023, with samples representative of the general population in Australia, Belgium, France, Germany, Japan, the Netherlands, Singapore, Spain, Sweden, the United Kingdom, and the United States, and representative of the online population of Brazil, India, Mexico, Poland, South Africa, and the United Arab Emirates. The 2024 Google / Ipsos Our Life with AI study was fielded online between September 17 and October 8, 2024, with samples representative of the general population in Australia, Belgium, Canada, France, Germany, Italy, Japan, the Netherlands, Poland, Singapore, South Korea, Spain, United Kingdom and the United States. The sample also included roughly 1,000 adults $18+$ representative of the online population of each in Brazil, Chile, India, Mexico, Nigeria, South Africa, and the United Arab Emirates.