How Disclosure Language Influences Retail Investor Attention When AI Writes for AI to Read
AI now sits on both sides of corporate disclosure.
On the production side, companies are increasingly using generative AI (GenAI) to draft earnings call scripts, SEC filings, news releases and other investor relations (IR) materials. Statistically significant evidence of AI-generated text has been detected across multiple disclosure types, with up to 4.5% of new content in 2024 attributable to GenAI. A 2025 survey by the National Investor Relations Institute (NIRI) found that nearly 60% of IR practitioners already use AI in their workflow, with universal expectation that adoption will increase significantly over the next three to five years.
In a parallel shift on the reception side, GenAI makes interpreting financial materials considerably quicker and easier, substantially lowering analytical costs for nonprofessional investors. Retail investors have therefore become highly active AI adopters, to the point where machine downloads and automated processing now dominate how disclosures are accessed.
An “AI Loop” is being created. Financial communications and IR teams are using AI as the co-author of disclosure content, while anticipating it has become the first reader. This recognition drives AI engine optimization (AEO), a tactical practice of structuring content to be favorably extracted and surfaced by AI tools. The result is self-reinforcing – AI writes in AI-friendly language for AI to interpret.
So what is AI’s language? Research across AI-assisted writing has identified a recognizable set of linguistic features that distinguish AI-generated text from human writing: higher readability and optimism, with more uniform and templated structure.
This emerging linguistic landscape is shaping financial communications. A recent study examined how the AI-style language in corporate disclosures influences retail investor attention. The analysis covered 36 Management's Discussion and Analysis (MD&A) sections from four major U.S. public companies across nine years of 10-K filings, spanning three phases of consumer-facing AI development: pre-AI baseline (2017-2019), early-AI expansion (2020-2022) and advanced-AI diffusion (2023-2025). More than 400,000 words were measured across six features in three dimensions: readability, sentiment tone (positive, negative and uncertain) and structural patterns (sentence length variation and phrase repetition). Corresponding to these indicators of AI-induced language style, retail investor attention was captured through two channels: search-based attention via Google Trends and discussion-based attention via Reddit activity.
Several findings stood out.
Readability, long the default measure of both disclosure quality and AI writing, had no impact on investor attention in either channel. Structural features similarly showed no consistent effect.
Sentiment tone produced the strongest attention response. For retail investors who lack institutional-level analytical resources, sentiment functions as the most accessible risk signal. Even when they parse all the complex information, emotional cues hinting at financial outlook remain the primary reading purpose. Specifically, when companies used less negative and less uncertain language relative to their historical norms, retail search and discussion around that stock both increased.
Most notably, the same linguistic profile produced divergent effects across the two attention channels.
Disclosures with stronger AI-style features attracted higher retail search throughout the observation window, which offers an intuitive confirmation of AEO: more structured and tonally restrained language is easier for algorithms to scan and rank, driving more search-based discovery.
In the discussion-based attention channel, higher AI-style content initially drew more engagement as well, but that advantage had evaporated within roughly six weeks of filing with the SEC. Lower AI-style disclosures overtook them and claimed a commanding lead in online communities, where conversations materially drive retail investors’ collective trading decisions and market impacts.
For IR professionals, no single content strategy can maximize both attention channels. When the priority is search-based visibility, raising algorithmic ranking and ensuring AI tools summarize the message accurately, sentiment-suppressed and streamlined language performs well. When the priority is discussion-based engagement, fueling public opinion and generating cost-effective organic traffic, it is essential to leverage emotionally rich expressions, varied structures and room for interpretation that sustain collective conversations.
Recognizing which objective each communication approach serves is where strategic IR in the AI Loop begins.
Editor’s Note: This article is adapted from graduate capstone research conducted at the New York University School of Professional Studies as part of the Master of Science in Public Relations and Corporate Communication program.
