A new national study reveals that large language models like ChatGPT are now widely used across all demographics in the U.S., with nearly half of users believing these AIs are more intelligent than themselves and many relying on them for major life decisions. (Source: Image by RR)

OpenAI’s ChatGPT Leads the Pack With 72% Adoption Among Users

A new study from the Imagining the Digital Future Center reveals widespread and rapidly growing adoption of large language models (LLMs) across the United States, with usage now evenly distributed across gender, income and education levels. ChatGPT remains the dominant platform, used by 72% of respondents, followed by Google’s Gemini (50%) and Microsoft’s Copilot (39%). Notably, 58% of users have tested multiple platforms, though only a small fraction (4%) personally pay for premium features. The overlap between language models and image generation is also strong—two-thirds of LLM users have tried tools like DALL·E and Midjourney, with nearly a third using them frequently.

LLMs, as noted in the-decoder.com, are primarily used for personal learning and planning (51%), rather than professional purposes (24%), though adoption varies significantly by occupation. Professionals in management, science, and education show high usage rates, but even retail and service workers report strong engagement (62%). Households with children and non-white users are especially active. The survey also found that 49% of users view LLMs as more intelligent than themselves, with women more likely than men to rate the models as significantly more intelligent. Voice-based interaction is a growing trend, especially among lower-income users.

Despite high satisfaction levels (76%), users report mixed experiences. Nearly a quarter admit to making serious mistakes due to incorrect information from LLMs, while others feel manipulated or overly reliant on the technology. Around one-third describe their use as “cheating” or express confusion. Still, LLMs are increasingly influencing critical life decisions: 41% of users consult them for career guidance, 37% for health advice, 28% for job changes, 25% for financial decisions and 18% for housing choices.

The study concludes that LLMs are reshaping how Americans learn, make decisions, and perceive their own capabilities. While the tools offer substantial benefits and convenience, they also raise ethical and cognitive concerns about dependency, accuracy, and the impact on human judgment. Conducted by SSRS between January 21–23, 2025, the research reflects a growing societal shift toward AI-driven assistance and underscores the urgent need for digital literacy and oversight.

read more at the-decoder.com