The Widespread Adoption of Large Language Model-Assisted Writing Across Society

The recent advances in large language models (LLMs) attracted significant public and policymaker interest in its adoption patterns. In this paper, we systematically analyze LLM-assisted writing across four domains—consumer complaints, corporate communications, job postings, and international organization press releases—from January 2022 to September 2024. Our dataset includes 687,241 consumer complaints, 537,413 corporate press releases, 304.3 million job postings, and 15,919 United Nations (UN) press releases. Using a robust population-level statistical framework, we find that LLM usage surged following the release of ChatGPT in November 2022. By late 2024, roughly 18% of financial consumer complaint text appears to be LLM-assisted, with adoption patterns spread broadly across regions and slightly higher in urban areas. For corporate press releases, up to 24% of the text is attributable to LLMs. In job postings, LLM-assisted writing accounts for just below 10% in small firms, and is even more common among younger firms. UN press releases also reflect this trend, with nearly 14% of content being generated or modified by LLMs. Although adoption climbed rapidly post-ChatGPT, growth appears to have stabilized by 2024, reflecting either saturation in LLM adoption or increasing subtlety of more advanced models. Our study shows the emergence of a new reality in which firms, consumers and even international organizations substantially rely on generative AI for communications.

Introduction. The emergence of large language models (LLMs) marked a significant moment in artificial intelligence, offering unprecedented capabilities in natural language processing and generation. This rapid proliferation of LLMs generated both excitement and concern. On one hand, LLMs have the potential to greatly enhance productivity; in the writing space specifically, it can democratize content creation (especially for non-native speakers). On the other hand, policymakers fear an erosion of trust, risks of biases and discrimination, and job displacement1–3; businesses worry about reliability and data privacy; academics debate the implications for research integrity and teaching4,5; and the public is concerned about misinformation, deepfakes, and authenticity6,7. Further complicating the discourse is the question of how LLMs may widen or potentially bridge socioeconomic gaps, given differential access to these advanced technologies. Although some early adoption stories or isolated examples have drawn significant media attention, and survey studies have explored LLM adoption from an individual user perspective8,9, there remains a lack of systematic evidence about the patterns and extent of LLM adoption across various diverse writing domains. While some previous work used commercial software to detect such patterns10,11, these studies often been constrained to single domains, relied on black-box commercial AI detectors, or analyzed relatively small datasets. To address this gap, we conduct the first large-scale, systematic analysis of LLM adoption patterns across consumer, firm and institution communications. Our analysis leverages a statistical framework validated in our previous work12 to quantify the prevalence of LLM-modified content. This framework has demonstrated superior robustness, transparency (and lower cost) compared to commercial AI content detectors12–14, allowing us to track adoption trajectories and uncover key demographic and organizational factors driving LLM integration. We focus on four domains where LLMs are likely to influence communication and decision-making: consumer complaints, corporate press releases, job postings, and United Nations press releases. Consumer complaints offer insight into user–business interactions and show how these technologies may extend beyond AI-powered customer service15. Corporate press releases reflect strategic organizational usage, as firms incorporate LLMs into their investor relations, public relations, and broader business communications. Job postings reveal how recruiters and human resource departments harness LLMs, shedding light on broader labor market trends. Finally, UN press releases showcase the growing institutional adoption of AI for regulatory, policy, and public outreach efforts.1 This comprehensive approach reveals several patterns. First, we observe a consistent trajectory across all the analyzed domains: rapid initial adoption following ChatGPT’s release, followed by a distinctive stabilizing trend highlighting widespread adoption. One of the remarkable results from our analysis is how similar adoption is between these diverse domains. By the end of the period we analyzed, in the financial dataset we estimate about 18% of the data was generated by LLM, around 24% in company press releases, up to 15% for young and small companies job postings, and 14% for international organizations. Second, we uncover some heterogeneity in adoption rates across geographic regions, demographic groups, and organizational characteristics. Third, we find that organizational age and size emerge as the most important predictor of differential adoption, with smaller and younger firms showing markedly higher utilization rates. Our findings provide crucial insights into the first wave of LLM integration across society, revealing how various socioeconomic and organizational factors shape technology adoption patterns. This understanding is essential for policymakers, business leaders, and researchers as they navigate the implications of AI integration across different sectors of society and work to ensure equitable access to and responsible deployment of these powerful new tools in the future.

Discussion. Our findings reveal widespread adoption of large language models across diverse writing domains, ranging consumers, firms and international organizations. This finding complements and extends our previous research that found widespread adoption across academic researchers.12 A consistent temporal pattern emerges from our data: after an initial lag of 3–4 months following the ChatGPT launch, there was a sharp surge in LLM usage, which then stabilized by late 2023 and remained steady through 2024. This trajectory deviates from traditional diffusion models that predict continuous and gradual growth, suggesting several possibilities. Early adopters may have already reached a saturation point within their domains, or domain-specific barriers (generally, these can range from costs of adoption, regulatory constraints, concerns over authenticity coupled with advances in users recognizing AI writing, etc.) that could be impeding further expansion. Alternatively, improvements in LLM sophistication may be rendering AI-generated content increasingly indistinguishable from human writing, complicating our ability to measure ongoing adoption. In the consumer complaint domain, the geographic and demographic patterns in LLM adoption present an intriguing departure from historical technology diffusion trends16–18 and technology acceptance model19,20, where technology adoption has generally been concentrated in urban areas, among higher-income groups, and populations with higher levels of educational attainment21,22. While the urban-rural digital divide seems to persist, our finding that areas with lower educational attainment showed modestly higher LLM adoption rates in consumer complaints suggests these tools may serve as equalizing tools in consumer advocacy. This finding aligns with survey evidence indicating that younger, less experienced workers may be more likely to use ChatGPT8. This democratization of access underscores the potentially transformative role LLMs could play in amplifying underserved voices. However, further study is needed to assess whether this increased adoption translates into more effective consumer outcomes. Corporate communication channels also demonstrated widespread but decelerating LLM integration. The plateauing adoption across platforms like Newswire, PRWeb, and PRNewswire raises important considerations about the balance between cost efficiency and authenticity. While LLMs may enable rapid, cost-effective content generation, overreliance on automated tools could compromise the nuance and credibility required in professional communications, potentially eroding trustworthiness23–25. Future research should explore how organizations navigate this trade-off and whether editorial interventions are employed to mitigate potential drawbacks.

In the recruitment process, small firms, particularly those founded after 2015, exhibited the fastest adoption of LLM-generated content. This trend suggests that younger, or companies closer to the technological frontier are leveraging LLMs to streamline hiring processes and reduce costs.5 While our study did not directly measure homogenization, prior research on the homogenization of LLM-generated content in academia12,13 suggests that similar effects could occur in job postings. This potential homogenization may inadvertently obscure critical distinctions between roles and organizations, potentially complicating job seekers’ decision-making. In fact, recent evidence has shown that while employers who leverage LLM to generate first draft of job post may receive more applications, they are less likely to make a hire27.