A comprehensive analysis shows how AI writing assistance has become widespread across various sectors - from consumer complaints to United Nations communications.
Researchers from Stanford University, the University of Washington, and Emory University analyzed more than 1.5 million texts across four professional sectors between January 2022 and September 2024. Their findings show that AI-assisted writing has firmly established itself in the studied fields.
The impact is particularly noticeable in press releases, where up to 24 percent of content now comes from generative AI systems or shows significant AI modification. The researchers examined over 537,000 press releases across three major platforms: Newswire, PRNewswire, and PRWeb.
ChatGPT's launch in November 2022 marked a turning point. After a 3-4 month adjustment period, AI usage increased dramatically. Newswire saw peak adoption at 24.3 percent in December 2023 before stabilizing around 23.8 percent. PRNewswire followed a similar pattern, peaking at 16.4 percent, while PRWeb showed comparable numbers. Technology and business sectors lead in adoption rates.
AI writing extends beyond corporate communications
The researchers identified a similar pattern of AI adoption in consumer complaints submitted to the U.S. Consumer Financial Protection Bureau (CFPB), where around 18% of complaints were AI-generated by late 2024.
Small and recently founded companies lead AI adoption in job advertisements. The study reveals significant differences based on when companies were established: Firms founded after 2015 use AI-generated texts for certain job postings at rates between 10% and 15%, while those founded between 2000 and 2015 show lower adoption, at just 5% to 10%. Companies established before 1980 rarely use AI assistance, with adoption rates below 5%.
Similar figures were found in UN press releases, where AI-generated texts accounted for slightly under 14%. UN teams in Latin America and the Caribbean show the highest usage at around 20 percent.
| Scope | Sample size | Measurement period | Highest adoption rate |
|---|---|---|---|
| Consumer complaints | 687,241 complaints | Jan 2022 - Aug 2024 | 17,7 % |
| Company press releases | 537,413 press releases | Jan 2022 - Sep 2024 | 24.3 % (Newswire) |
| Job advertisements | 304.3 million job advertisements | Jan 2021 - Oct 2023 | Up to 15 % for small, young companies |
| UN press releases | 15,919 press releases | Jan 2019 - Sep 2024 | 13,7 % |
While the researchers see this trend as potentially democratizing professional writing, they caution about possible drawbacks like text standardization and loss of authenticity.
Actual AI usage likely exceeds study results
The researchers suspect that actual AI adoption rates are higher than their analysis suggests. It likely missed heavily human-edited content and text from advanced AI models that closely mimic human writing. The study also didn't examine other potential AI writing use cases, such as social media content creation.
An earlier study by the same research team found increasing AI presence in academic research and scientific publications. The researchers developed a specialized statistical method using maximum likelihood estimation (MLE) to detect AI content. This approach compares known human and AI-generated reference texts to estimate AI content proportions in larger text collections. They say this method offers more robust, accurate, and efficient detection than traditional approaches by analyzing statistical patterns at an aggregate level rather than classifying individual texts.
