Does chatbot language style actually shape how much we trust it?

The paper's central claim is that generative AI chatbots "leverage natural language signifiers of expertise and intelligence to influence users' perception of their trustworthiness." The authors reach this through procedural rhetoric, which reads a system by the procedures it makes available and the values those procedures carry. Their target is the set of procedures around information seeking. They argue that these are shaped by users' "values, experiences, and expectations" and are now being altered by "the emerging turn towards AI 'experts' and authority."

The discussion names the shift in one sentence. Users are "shifting away from the search-and-recall method wherein the onus is on the user to find, consume, and comprehend information of varying levels of difficulty, relevance, and reliability." In its place, the chatbot is given "the authority to find, filter, and assemble information." Trust is therefore not a side effect of good answers. It attaches to the register of the answer, and the register is what licenses the handoff of work that used to be the user's. The introduction frames this against a history in which routine information seeking meant consulting a reputable source such as the local newspaper, a family doctor, or a trusted neighbor. Chatbots now occupy roles from scheduler to tutor, coach, counsellor, and confidante, so the expert seat is being filled by a system.

Against the nearest notes, this paper supplies a rhetorical account of something the library already holds as an empirical finding. Does conversational style actually make AI more trustworthy? shows in a focus group that conversational style drives trust. This paper says why that style works: it reads as expertise, and expertise is the signal people use to decide whom to consult. It also sharpens How do we learn to read AI-generated text critically?. The missing interpretive posture matters most when the source speaks in the voice of an authority whose credentials we normally check. It sits beside Does polished AI output trick audiences into trusting it?, where presentation stands in for judgment, here extended from finished artifacts to the conversational voice itself. It contrasts with How does LLM-mediated search change what expertise requires?, which keeps the burden of steering on the user, while this paper describes users letting that burden go.

The excerpt is silent on method. It gives no sample, no study design, and no measured effect of expertise signifiers on trust. The procedural-rhetoric analysis lives in the original journal article, and the introduction breaks off mid-citation. The discussion poses its central questions and does not answer them: what it means to trust a chatbot with information seeking, and what happens when decisions are outsourced to something that "cannot have true experience or direct knowledge in the world it claims to understand." What follows at this strength is a framing. The language of expertise is a design surface that can raise perceived trustworthiness independent of the system's grounding in the world, and the shift away from search-and-recall moves the user's evaluative work onto that surface. Whether users actually calibrate to it is not shown here.