Research opportunity · Persuasion, Intermediation & the Information Order
Does weeks of everyday AI-answer use make people check sources less?
Earlier version (v1)
This is RO-LD10-04 as it was published at v1, kept unchanged. The current version is v2; its changes are listed there.
Question
When people are randomly assigned to search with AI answers as the default, or without them, for eight to twelve weeks of their normal information seeking, how do their source-checking behaviour, trust in AI answers, ability to spot wrong answers and reliance on AI for news and health questions change over time, and do early effects fade, persist or grow?
Not yet measuredOne PhD chapter or paperNeeds a budget (medium or higher)Partly addressed
Why it matters
The worry about AI intermediation is less about any single answer than about habits: whether repeated reliance gradually replaces checking and comparing sources. The field experiments found last one to two weeks, the longitudinal panels found are observational, and the longest randomized chatbot study found measured well-being rather than information habits. A study that tracks behaviour over months is needed before anyone can say whether everyday AI answers build epistemic dependence.
The debate lists this among what we still do not know: repeated everyday-assistant effects on epistemic dependence
Evidence that motivates it
- In a preregistered one-week U.S. field experiment, forced Google AI Mode reduced external click-through by 18.8 percentage points and lowered trust in information on Google.
A one-week forced AI Mode experiment lowered external click-through and trust; the claim does not license long-run adoption effects, and longer natural-adoption studies would change it. - In a preregistered U.S. experiment, citations increased trust in AI-search answers even when incorrect or hallucinated, while uncertainty highlighting reduced trust and sharing.
In a single-session experiment, citations raised trust in AI-search answers even when wrong, while uncertainty highlighting reduced trust; whether this carries over to field reliance is not established. - In one UK experiment, about 36%–42% of GPT-4o's immediate attitude effect remained at a one-month recontact.
The only persistence estimate in the debate is a one-month recontact for persuasion, which shows how short current follow-up horizons are; it does not license claims about durable change.
A possible design
Recruit about 1,000 adults who use search daily. After a two-week baseline of logging, randomize them for 8 to 12 weeks into (a) search with AI answers hidden, (b) normal search with AI answers, and (c) AI answers with uncertainty highlighting added by the extension. Log searches, AI-answer exposure, clicks to sources, use of multiple sources per query and chatbot visits. Every two weeks embed short tasks with known answers, some with planted errors in AI-style summaries, to measure detection of wrong answers. Survey trust, perceived need to verify and self-reported reliance at baseline, mid-point, end and a six-week follow-up after the condition ends. Estimate effects by week to see whether they accumulate or fade, and compare heavy and light baseline users.
Data needed
A browser-extension logging and display tool (desktop, with a mobile pilot if feasible); participant payments for 10 to 16 weeks; validated trust and reliance scales; a bank of verification tasks with planted errors; ethics approval for log collection and deception debrief.
Feasibility
One PhD chapter or paper. Medium: about US$20,000 to 50,000 in participant payments for a three-month panel with attrition incentives; engineering time can reuse open-source tools from existing search field experiments. Attrition over three months is the main threat; budget for retention bonuses and preregister attrition analyses. Hiding AI answers for months may not be possible if platforms change their interface, so monitor the manipulation weekly.
Has this been done?
Partly done already: related work covers some of this question (first-pass screen of 12 October 2026; not yet checked in depth).
A randomized field experiment with AI answers in search exists for seven days (Wang et al.), a four-week randomized chatbot experiment measured loneliness and emotional dependence rather than information verification (Fang et al.), a year-long six-wave panel run from Germany (IWM Tübingen) studies trust and willingness to delegate tasks to AI observationally, and a four-week qualitative study followed trust in Snapchat's My AI with 27 people. No randomized study of several months that measures source-checking behaviour and detection of wrong answers in everyday search was found.
How strong this evidence is
- Who searched and checked: an independent AI reviewer with fresh context (not a person), 12 October 2026: every nearest-work citation checked against its source page; corrections applied. Every check so far was done by AI; no person or domain expert has checked it yet.
- What was read: none of the 4 nearest works below is recorded as judged from an abstract only. Not every entry says how much of the work was read.
- Where the search could not reach: Screened by web search on the date shown. ScienceDirect and Taylor & Francis article pages could not be opened, so two relevant panel studies (the IWM article and a two-wave cross-cultural panel on trust in AI news in Behaviour & Information Technology) were seen only through listings. Google Scholar was not searched. Absence of a hit is not proof that the work does not exist.
- Before building a project on it, run your own search in a scholarly database such as Scopus, Web of Science, Google Scholar or SSRN.
Nearest work found
- Wang, Gleason, Bart, Wilson & Metaxa, "AI in Search Reduces Publisher Referrals Without Improving User Experience: Experimental Evidence," arXiv:2608.18352 (18 Aug 2026)
Preregistered US field experiment, 1,100 participants, three-day baseline then seven days in condition; measures click-through, user experience and trust. Its limitations note that longer exposure periods may show different effects; it does not track verification habits over time. - Fang, Liu, Danry, Lee, Chan, Pataranutaporn, Maes, Phang, Lampe, Ahmad & Agarwal, "How AI and Human Behaviors Shape Psychosocial Effects of Extended Chatbot Use: A Longitudinal Randomized Controlled Study," arXiv:2503.17473 (v2, Oct 2025)
Four-week randomized study of 981 participants comparing chatbot modes (text, neutral voice, engaging voice) and conversation types; outcomes are loneliness, socializing with people, emotional dependence and problematic use. The abstract does not mention measuring information verification or reliance for knowledge. - Ehrhardt, Ivan & Klein, "From perception to adoption: A longitudinal study of trust and delegation in everyday AI use," Computers in Human Behavior: Artificial Humans 8, 100325 (May 2026), from the IWM Tübingen project "A longitudinal study on the perceptions and dynamics of human-AI interaction" (August 2024 to April 2026)
The project page describes a six-wave, one-year observational panel measuring trust, reliance, credibility and willingness to delegate tasks to AI. Bibliographic details were confirmed through the DOI record and the project page (https://iwm-tuebingen.de/en/research/projects/hmc_ai_trends); the article text was not reachable. Survey-based and not randomized; no logged search behaviour. - Landerberg, Flatmo & Said, "Trust as a Situated User State in Social LLM-Based Chatbots: A Longitudinal Study of Snapchat's My AI," arXiv:2604.22417 (24 Apr 2026)
Four-week qualitative survey study of 27 users describing how trust and reliance shift over time. Small, qualitative and about a social chatbot, not search.
Search queries used
longitudinal study repeated chatbot use reliance trust epistemic dependence weeks randomized assignment everyday usepanel survey generative AI use trust in information over time news verification habits waves 2025 2026multi-week field experiment AI assistant for information seeking effects on verification behavior overreliance diary"From perception to adoption: A longitudinal study of trust and delegation in everyday AI use"
Provenance
RO-LD10-04 v1. Drafted by Claude (SocioLens Project Lead), AI-drafted from the LD-10 debate on 2026-10-12. Checked by an independent AI reviewer with fresh context (not a person), 12 October 2026: every nearest-work citation checked against its source page; corrections applied. Not yet reviewed by a domain expert. This is a research hypothesis, not a finding.