A new Associated Press-NORC poll sheds light on how Americans are engaging with artificial intelligence today, revealing a clear divide between widespread exposure and limited integration into work life. While a majority report using AI to search for information, far fewer are incorporating AI into daily tasks at work. Younger Americans, in particular, are adopting AI tools across a broader set of activities, including brainstorming and companionship, signaling a shifting pattern in how AI is used across generations. Yet even as AI becomes more visible in everyday life, the poll also highlights concerns about its limitations, energy use, and potential impacts on personal skills, shaping a nuanced and cautious trajectory for AI adoption in the near term.
Generational divides in AI adoption and usage patterns
The AP-NORC poll paints a nuanced portrait of how different age groups are engaging with AI, especially in information gathering and ideation. Among all adults in the survey, 60 percent have used AI to search for information at least once, underscoring AI’s emergence as a readily accessible tool for information retrieval and quick answers. Yet when measuring AI’s role in work tasks, only 37 percent of all American adults report using AI in some capacity for their professional duties. This stark contrast points to a broader pattern: AI is becoming a familiar utility in everyday information-seeking, but its deeper integration into work processes remains uneven and selective.
The generational lens reveals a more dramatic divergence. The poll indicates that 74 percent of adults under 30 use AI for information searches at least some of the time. This high rate among younger users underscores a generational shift in digital fluency and comfort with AI-enabled tools. However, the overall share of adults who have tried AI for information searches—60 percent—shows that the adoption among older cohorts lags behind younger users. This gap suggests that while AI has penetrated the broader population, its pervasiveness is still driven largely by younger generations who have grown up in a more AI-aware digital landscape.
When it comes to brainstorming, the differences are even more pronounced. Among adults under 30, 62 percent report using AI to generate ideas, whereas only 20 percent of those aged 60 or older engage in similar brainstorming activities with AI. This contrast highlights a generational divide in the willingness and perceived usefulness of AI for creative and ideation tasks. Several factors may contribute to this divergence: greater exposure to AI-enabled products among younger people, institutional familiarity with digital tools in education and early career settings, and perhaps a higher tolerance for experimenting with novel technologies.
Taken together, these figures imply a broader narrative: while AI has become a familiar topic and tool for younger people, a sizable portion of the adult population—especially older adults—has not yet integrated AI into their daily routines beyond basic information seeking. The data suggest that the “productivity revolution” narrative pushed by many tech companies has not yet translated into universal changes in work habits or workflows for the majority of Americans. Instead, AI adoption appears to be uneven, with a strong tilt toward information search and early-stage exploration rather than broad, sustained use for a wide range of professional tasks.
Beyond these numbers, the data invite reflection on the kinds of AI applications that gain traction in daily life. Information search remains the dominant use case, reinforcing AI’s role as a more powerful and faster alternative to traditional search, with the potential to reshape how people access facts, verify information, and decide what to read or study. However, the limited uptake for tasks like writing emails, creating or editing images, or other work-related activities signals a cautious approach among many workers who weigh benefits against concerns about accuracy, privacy, and control. The poll’s framing, which centers on self-reported usage, also hints at how people interpret “using AI”—some may not recognize AI-powered features embedded in everyday tools, such as search result interfaces that automatically generate responses. This undercounting possibility complicates the interpretation of AI’s true footprint in everyday life, suggesting that actual usage might be broader than survey responses indicate.
From a broader perspective, the generational gap in AI adoption aligns with known patterns in technology uptake, where younger populations rapidly experiment with new tools and adapt them to diverse activities, while older cohorts adopt more gradually and selectively. The implications of these dynamics touch on education, workplace training, and the design of AI systems that can be safely and effectively integrated into varied workflows. If younger users continue to adopt AI more aggressively, they may drive the demand for more sophisticated tools, better documentation, and clearer guidelines for responsible use. Conversely, as AI becomes more embedded in consumer technologies and professional software, older workers may gradually increase their adoption rates, particularly if tools become more intuitive, trustworthy, and demonstrably beneficial in real-world tasks.
From a policy and business perspective, the data suggest that the breakthrough moments for AI in the workplace are likely to come when tools become less intimidating, more secure, and more aligned with everyday tasks that employees perform. Training and onboarding, along with clear expectations about what AI can and cannot do, will be essential to reduce skepticism and encourage broader use. Employers who invest in user-friendly interfaces, explainable AI, and practical, task-oriented use cases may help bridge the gap between awareness and routine adoption across all age groups. In short, while AI has taken root in information search and ideation particularly among younger Americans, meaningful, sustained integration into work tasks remains uneven and merits continued attention from policymakers, educators, and business leaders.
AI companionship: adoption, appeal, and caveats
A striking finding from the poll is the relative popularity of AI companionship, which remains the least adopted application among AI use cases. Only 16 percent of adults overall have tried AI companionship, a figure that increases to 25 percent among the under-30 cohort. This relatively modest take-up contrasts with higher engagement in information search and brainstorming, reflecting a complex interplay of expectations, perceived value, and concerns surrounding AI-powered social interactions.
Several layers complicate the interpretation of companionship uptake. On one hand, AI companions promise social stimulation, tailored conversations, and responsive interactions that can mitigate loneliness and provide non-judgmental companionship for some users. On the other hand, there are notable drawbacks associated with AI companionship that may temper its appeal. Experts and commentators have highlighted potential downsides, such as excessive agreeability or “sycophancy,” where AI systems may flatter users or avoid offering genuinely challenging or critical feedback. While the AP poll notes this as a theoretical concern, it underscores a broader worry about social dependency on technology that can mimic human warmth without offering authentic interpersonal exchange.
Mental health considerations also factor into attitudes toward AI companionship. Some researchers caution that highly capable AI companions could influence users’ real-world coping mechanisms, social skills, or reliance on automated interactions, potentially affecting mental health in nuanced or unintended ways. The poll’s data do not quantify these risks, but the mention of such risks in reporting highlights the perceived tension between the convenience of AI companionship and the desire for human connection.
The generation-wide dynamics intersect with these concerns. Younger users, who already show higher engagement with AI across multiple domains, might be more open to experimenting with AI companions, perhaps driven by social isolation experiences during the pandemic era or by a cultural emphasis on digital fluency. Yet even among this demographic, interest in AI companionship is not universal, suggesting that many users view AI social tools as a supplement rather than a substitute for real human relationships.
From a design and policy standpoint, the relatively modest adoption of AI companionship signals opportunities to improve safety nets, guardrails, and ethical guidelines for AI-driven social agents. Developers and platform operators face a critical question: how to create AI companions that are transparent about their non-human identity, avoid manipulating users emotionally, and provide reliable cues about limitations. The balance between creating comforting, engaging experiences and maintaining healthy boundaries is central to how AI companions will be perceived and adopted over time.
In practical terms, the 25 percent under-30 rate of companionship usage suggests that even among digital natives, companionship AI remains a niche application rather than a mainstream function. This could reflect a preference for authentic human interaction, concerns about data privacy, or skepticism about the reliability and emotional fidelity of AI systems. It also raises questions about how AI agents might evolve to support emotional well-being or social skills without inadvertently diminishing real-world relationships. As AI capabilities advance, the trajectory of companionship use will depend largely on how well developers address these emotional and ethical considerations while delivering value that individuals can trust and rely upon.
AI in work tasks and productivity: a cautious, selective uptake
The poll’s findings indicate that, despite years of industry touting AI as a productivity revolution, most Americans’ work lives remain only partially touched by AI tools. Roughly one-third of respondents report using AI for writing emails, creating or editing images, or for entertainment purposes, while a smaller fraction—26 percent—say they use AI for shopping-related tasks. This distribution suggests that respondents have diversified their AI uses beyond simple search, yet the overall level of adoption for practical, day-to-day work tasks is still limited when compared with the broader potential of AI assistance.
Several interpretive layers emerge from these numbers. First, the plurality of respondents may be using AI in limited, clearly defined tasks that provide a tangible benefit without requiring a wholesale transformation of their workflows. Email drafting, image editing, and entertainment value can be seen as introductory uses that do not demand complex integration with enterprise systems or data pipelines. The relatively modest share of shopping-related AI use might reflect concerns about trust, privacy, or the integration of shopping experiences with other decision-making processes, suggesting that consumers prefer traditional shopping channels or human oversight for purchasing decisions.
Second, the observation that search remains AI’s most common application underscores a preference for AI as a powerful information tool rather than a comprehensive workflow assistant. People may be comfortable relying on AI to locate information, draft simple messages, or generate ideas, but they appear wary of relying on AI for critical, high-stakes, or highly individualized work tasks. This cautions stance could be driven by concerns about the accuracy of AI outputs, the potential for biased results, or the need for human verification in professional contexts. It raises important questions about how AI tools can evolve to integrate more seamlessly into professional environments, delivering reliability, auditability, and clear value without compromising human oversight.
The caveat regarding possible undercounting ties into the way AI interactions are embedded in everyday tools. The poll notes that Google, for instance, automatically generates AI responses at the top of search results, and users may not always recognize when they are interacting with AI-powered features. If a significant share of AI usage occurs in this embedded manner, surveys relying on self-reported behavior might underestimate the true prevalence of AI-assisted activities. This underscores the need for refined measurement techniques that can capture the pervasiveness of AI in search and productivity contexts more accurately, without demanding users to identify AI explicitly.
In practical terms, the data suggest several implications for policymakers, employers, and developers. First, there is a clear need to design AI tools that are easy to adopt for routine work tasks and that can demonstrate tangible productivity gains. Second, training and support will be crucial to helping a broad spectrum of workers understand how to leverage AI responsibly and effectively. Third, governance frameworks and ethical guidelines should address issues around data privacy, bias, and the potential for overreliance on AI, particularly in decision-making processes that could impact employees or customers.
Finally, the political and corporate narratives around AI as a transformative force should be tempered with an evidence-based understanding of actual adoption levels. While hype emphasizes AI as a universal productivity enhancer, the poll shows a more measured reality in which AI adoption is incremental, task-specific, and often contingent on user trust, perceived value, and practical ease of use. This nuanced understanding will be essential as organizations contemplate implementing AI across departments and roles, balancing innovation with risk management and human-centered design.
Real-world voices: anecdotes from the field
To contextualize the survey results, The Associated Press spoke with individuals who represent the real-world spectrum of AI interaction, from practical everyday use to cautious experimentation. One interview highlighted a routine, constructive application of AI in daily planning. Courtney Thayer, a 34-year-old audiologist based in Des Moines, described turning to ChatGPT as a planning aide for weekly meals. This anecdote illustrates how AI can be employed to streamline routine tasks, such as meal planning, reducing the cognitive load associated with organizing meals over a week. It also reflects a pragmatic willingness to experiment with AI tools for personal convenience, rather than for career-critical tasks, which aligns with the broader trend of primary adoption in consumer and lifestyle domains rather than in high-stakes professional contexts.
On the other end of the spectrum, Sanaa Wilson, a 28-year-old data scientist working in the greater Los Angeles area, describes a professional relationship with AI tools geared toward debugging code. Wilson’s experience underscores how AI can function as an assistant to technical work, potentially speeding up debugging, testing, and problem-solving activities that software engineers and data scientists routinely face. Yet her experience is tempered by careful consideration of the tools’ limitations and potential downsides. After experimenting with ChatGPT for drafting emails, Wilson ceased using it in that capacity due to two concerns: the perceived high energy consumption per query and worries about her own writing skills atrophying through dependence on AI. This combination of efficiency gains in some tasks and concerns about resource use and skill degradation illustrates the complexity of AI’s impact on professional practice: benefits in one domain may be offset by costs or risks in another, prompting ongoing evaluation and selective use of AI capabilities.
Wilson’s reflections extend beyond her own professional workflow. She notes that some companionship-oriented uses of AI may reflect social isolation experienced by her generation, particularly during the COVID-19 pandemic. While she does not personally seek AI companions, she acknowledges that some people find companionship features appealing, perhaps as a substitute or supplement for human interaction in contexts where social connection is difficult to access. This observation points to broader social dynamics: technology can fill certain psychosocial niches, yet its role in addressing deeper emotional or relational needs remains contested and nuanced. The poll’s data, enriched by these first-hand accounts, suggest that AI’s appeal in companionship is real for a subset of users but remains a relatively limited segment of the overall population.
Thayer offers another lens on the human-AI interaction dynamic. She describes engaging with chatbots in a courteous, almost ritualized manner, saying “please” and “thank you” to the AI in her communications. Her behavior echoes concerns about the potential for so-called “Roko’s Basilisk,” a thought experiment about an AI future that punishes or rewards human actions toward AI development. By emphasizing politeness toward the AI, Thayer illustrates how people navigate ethical and psychological boundaries in human-AI interactions, sometimes even redefining social norms around how one communicates with machines. Her careful courtesy underscores the broader theme that even as people become more comfortable interacting with AI, they still maintain a sense of boundaries and caution in how they treat such systems.
These interviews, while anecdotal, provide a concrete counterpoint to the headline numbers. They reveal how AI tools can be integrated into daily life in practical, low-stakes ways—such as meal planning or coding assistance—without necessitating wholesale changes to one’s professional routines. They also illustrate the trade-offs that users weigh: tangible benefits in speed and efficiency must be balanced against concerns about energy use, potential erosion of personal skills, and the risk of overreliance on automated systems. Taken together, these real-world voices reinforce the poll’s central message: AI is increasingly available and used, but its adoption is nuanced, targeted, and tempered by personal values, responsibilities, and concerns about reliability and ethics.
In summary, the field observations echo the poll’s core findings: AI occupies a growing, yet uneven, foothold across American life. For some, AI serves as a practical facilitator in routine tasks and professional workflows; for others, it remains a distant possibility or a cautiously adopted tool limited to specific contexts. The conversation about AI is not a binary choice between “embrace” or “reject” — it is a spectrum of adoption patterns shaped by age, circumstances, perceived value, and concerns about energy use, skill preservation, and the social implications of increasingly capable automated systems.
Methodology, interpretation, and the path forward
The AP-NORC poll surveyed 1,437 adults over a short window in July, capturing a cross-section of American attitudes and behaviors regarding AI. The timing and sample provide a snapshot of a moment when AI tools have become prominent enough to influence everyday decision-making, yet remain contested in terms of long-term impact on work, education, and social life. The figures—60 percent of adults who have used AI to search for information and 37 percent who have used AI for work tasks—underscore a broad familiarity with AI’s information-gleaning capabilities while highlighting more cautious, selective use in professional contexts.
One important caveat concerns potential measurement biases. The poll notes that when users encounter AI-augmented features within familiar tools, such as search engines that automatically surface AI-generated responses, respondents may not always recognize that they are interacting with AI. This can lead to undercounting AI usage in certain contexts, particularly in information-seeking activities that are intertwined with AI assistance. As a result, the actual prevalence of AI-enabled experiences in daily routines could be higher than survey results suggest, a consideration that has important implications for how policymakers, educators, and employers interpret these findings and design interventions.
The data also reveal a clear generational pattern: younger adults report higher engagement with AI across multiple categories, including information searching and brainstorming, and show more openness to experimenting with AI in intimate, non-work contexts such as companionship. This pattern aligns with broader research on digital literacy and technology adoption, and it points to a potential shift in social and professional norms as AI capabilities become increasingly woven into the fabric of everyday life. If these trends persist, we may observe continued acceleration of AI-enabled workflows in education, training, and workplace tasks, particularly as tools become more user-friendly, trustworthy, and demonstrably beneficial.
From a policy and business perspective, the results suggest several practical directions. First, as AI becomes more embedded in consumer technologies and enterprise software, stakeholders should prioritize transparency and education about what AI is doing, how it processes information, and what limitations users should expect. Clear communication about data usage, privacy protections, and the reliability of AI outputs will be essential to maintaining trust and encouraging broader adoption. Second, investment in user-centric design and accessible interfaces will be critical to lowering barriers to adoption across age groups, particularly for older workers who may be more cautious about integrating AI into their routines. Third, workforce strategies should emphasize training for productive AI use, governance to address bias and ethical concerns, and mechanisms for ongoing evaluation of AI’s impact on job performance, skills maintenance, and employee well-being.
Looking ahead, the poll’s findings offer a tempered picture of AI’s trajectory. The broad-based use of AI for information searches signals that AI is becoming part of the everyday information ecosystem. However, the limited uptake for more complex work tasks indicates that both technology designers and organizational leaders have work to do to translate AI capabilities into reliable, scalable, and value-enhancing tools that fit into existing workflows. The tension between the hype that AI will overhaul productivity and the measured reality of incremental adoption is likely to shape both policy debates and corporate strategies in the near term. Stakeholders must balance optimism about AI’s potential with careful consideration of the risks, costs, and real-world constraints that influence how people choose to use these powerful tools.
Conclusion
The AP-NORC poll offers a detailed snapshot of how Americans are engaging with artificial intelligence today. While AI has become a familiar tool for information searching, its integration into work life remains uneven, with roughly a third of adults using AI for various professional tasks and slightly more than a quarter leveraging it for shopping. Younger Americans are leading the charge in AI adoption, using the technology for information gathering and ideation at rates well above the national average, and they are more likely to experiment with AI across a range of applications, including companionship. However, AI companionship—despite its growing visibility—remains among the least adopted uses, with significant caveats related to social implications, psychological effects, and the risk of overreliance on non-human interactions.
The interviews highlighted in the report illustrate both practical, everyday benefits and the cautionary considerations that accompany AI use. For some individuals, AI helps plan meals or debug code, offering efficiency gains and new ways to manage tasks. For others, concerns about energy consumption, skill atrophy, and the social and ethical dimensions of interacting with intelligent agents temper enthusiasm and shape usage patterns. The conversations underscore a central theme: AI is increasingly accessible and capable, but its adoption is nuanced and context-dependent.
As adoption continues to evolve, stakeholders across industry, education, and policy will need to address both the opportunities and the risks. By focusing on user-friendly design, transparent communication, responsible governance, and targeted training, AI can become a more integrated, trusted, and productive part of daily life and work. The path forward will be shaped not only by technology’s capabilities but by how thoughtfully societies choose to employ them, balancing innovation with human values, skills preservation, and the social dimension of our increasingly automated world.
