Original title: Investigating the Impacts of Generative AI on Information Seeking
Alexi Orchard, Shannon Lodoen · September 2026
People now turn to artificial intelligence when looking for answers, but that creates a real problem. We do not fully know how these tools shape what we believe. Researchers studied this issue by looking at how AI chatbots speak. They examined the way these programs use words that sound smart and experienced. This wording works by mimicking human knowledge, which convinces people to trust the bot. In this paper, the authors expand on that study to build a shared vocabulary across different fields. This common language will help people describe, study, and reduce the risks of using AI in communication. They conclude that we must understand how personal values guide our search for answers, especially as AI becomes an authority.
You are buying a used car from a dealer who speaks with authority, uses technical terms, and flashes a certificate. The confident pitch makes you trust the car's quality, but you later discover the certificate is fake and the dealer has no real expertise. In this paper, AI chatbots use similar language tricks—confident phrasing and expert-like vocabulary—to appear trustworthy. The researchers analyze how these design decisions shape your willingness to rely on the AI's information.
When you ask a chatbot for information, it often sounds confident and knowledgeable—but that confidence is designed to persuade you, not guarantee truth. This matters because people increasingly rely on AI for research, health advice, or news, and mistaking a convincing tone for real expertise can lead to bad decisions or misplaced trust.
The abstract does not state empirical limitations, but as a conceptual and rhetorical analysis, this work does not provide experimental evidence for how users actually respond to these language cues. It proposes a vocabulary for discussion but does not test its effectiveness in real-world settings.
Ask anything about this paper. Answers come from its abstract, and say so when the abstract doesn't cover your question.
Investigating the Impacts of Generative AI on Information Seeking — Alexi Orchard, Shannon Lodoen
Abstract licence: CC0
Explanation generated by Synthesis from the abstract — it is not the original paper. This is a preprint, so it hasn't necessarily been peer reviewed.