How AI Chatbots Are Reshaping Everyday Language: Insights from Recent Research
Since its launch in late 2022, ChatGPT has rapidly transformed the way people interact with technology. Within just five days, this AI-powered conversational agent attracted over one million users. Today, surveys reveal that more than 50% of adults under 30 and nearly half of those aged 30 to 49 have engaged with ChatGPT, highlighting its widespread adoption and influence on communication habits.
Examining the Linguistic Impact of AI on Spoken English
A groundbreaking study titled “Model Misalignment and Language Change: Traces of AI-Associated Language in Unscripted Spoken English” offers the first peer-reviewed analysis of how large language models (LLMs) like ChatGPT are affecting human speech patterns. Accepted for presentation at the upcoming eighth Conference on AI, Ethics, and Society, this research investigates whether shifts in everyday language stem from users mimicking AI suggestions or if AI is actively reshaping the human linguistic system.
Key Findings: AI Buzzwords Entering Daily Conversation
Assistant Professor Tom Juzek, who led the study, explains, “Our analysis of lexical trends before and after ChatGPT’s debut reveals a notable alignment between human word choices and AI-generated language, especially with the rise of AI-related buzzwords.” Unlike spikes in word usage driven by external events-such as the surge in “Omicron” during the pandemic-this research identifies a distinct increase in terms like “delve” and “intricate” within educational and academic discussions, likely influenced by AI’s conversational style.
Juzek emphasizes the unprecedented scale of this linguistic shift: “The rapid and widespread adoption of certain formal and precise words in spontaneous speech is remarkable compared to historical language trends. Many of these terms are typically overused by AI, suggesting a direct connection.”
Words on the Rise: Formal Vocabulary Gains Ground in Speech
Words such as “surpass,” “boast,” “meticulous,” “strategically,” and “garner” have all experienced significant growth in spoken language frequency since ChatGPT’s release. These terms, traditionally more common in formal writing, are now increasingly present in casual conversation. The study found that nearly 75% of these targeted words showed increased usage, with some more than doubling in occurrence.
Methodology: Analyzing Millions of Words from Natural Speech
The research team, including computer science students Bryce Anderson and Riley Galpin, examined over 22 million words from unscripted spoken sources such as science and technology podcasts. Their analysis revealed a clear uptick in AI-associated vocabulary post-ChatGPT, while synonymous words without AI ties did not show similar increases. For instance, “underscore” surged in usage, whereas its synonym “accentuate” remained stable. This pattern suggests that AI-generated language is subtly permeating everyday speech, a phenomenon the researchers term the “seep-in effect.”
Broader Implications: Language as a Reflection of AI’s Societal Influence
Lead author Anderson reflects on the significance of these findings: “Language is humanity’s most powerful communication tool. Understanding how AI shapes this medium is crucial, especially as AI’s impact extends beyond tools to influence cultural and social norms, including how we express ourselves.”
This study builds on the team’s earlier work exploring AI’s transformative effects on scientific writing, which uncovered substantial structural changes in academic English. While Galpin presented those findings at the 38th International Florida AI Research Society Conference, Anderson and Juzek will share the new insights on spoken language at the AIES Conference this October.
Ethical Considerations and Future Directions
Galpin highlights the ethical dimensions of this research: “As LLMs increasingly shape human language, it raises critical questions about model biases and misalignments potentially influencing human behavior and communication patterns.”
Conducted through Florida State University’s Undergraduate Research Opportunity Program, this project exemplifies how high-achieving students can collaborate with faculty to explore cutting-edge topics.
Looking ahead, Juzek notes, “A vital question remains: Is AI merely accelerating existing language evolution trends, or is it actively driving new linguistic changes? Further research will be essential to unravel these dynamics.”
For more information on ongoing linguistic and AI research, visit the Department of Modern Languages and Linguistics at mll.fsu.edu.
