TL;DR: When searching for colorectal cancer symptom information, people consistently trust and prefer traditional search engines like Google over AI chatbots like ChatGPT . Trust emerged as the dominant factor driving whether people would actually use a search tool for health decisions.
A randomized controlled trial of 762 Texas adults examined how people evaluate different tools for finding colorectal cancer information. Researchers showed participants either Google search results or ChatGPT responses about colorectal cancer symptoms, presented in two scenarios: one depicting serious symptoms (blood in stool) and another describing minor symptoms (occasional gas). The study then measured acceptance using established frameworks like the Technology Acceptance Model II, which captures whether people find something useful, easy to use, and trustworthy.
The results were unambiguous: Google outperformed AI across every measured dimension. Participants rated traditional search engines as significantly more useful, easier to navigate, and more trustworthy than AI-generated responses. This gap persisted regardless of whether participants viewed serious or minor symptom descriptions. The consistency of these preferences points to something deeper than simple familiarity. Google's advantage appeared tied to transparent source attribution - people could see where information came from and evaluate individual sources. ChatGPT, by contrast, generated synthesized responses without explicit source links, which users apparently experienced as less trustworthy and more cognitively demanding to evaluate.
Trust emerged as the single strongest predictor of whether someone would actually use a search tool for health decisions. This finding matters because it suggests the gap between Google and AI won't close simply by improving AI accuracy. If people don't trust a source, better information won't overcome that skepticism. Notably, the severity of symptoms didn't shift preferences. People didn't turn to different tools when worried versus mildly concerned. Their technology choices remained stable, suggesting technology preference operates somewhat independently from health anxiety levels.
The authors note an important implication: as AI becomes more integrated into health information ecosystems, public response to cancer symptom information may shift. If people are less likely to trust AI-generated health advice, they might delay care-seeking or discount information that actually could prompt appropriate screening. For healthcare systems and communicators, this suggests a gap between where AI capability stands and where public readiness stands.
If you're searching for health information online, particularly about symptoms, pay attention to source transparency. Google's format explicitly shows you where information originates, allowing you to evaluate Mayo Clinic differently than a personal blog. That transparency appears to be why people trust it more. When using AI tools for health questions, treat them as a starting point rather than a final answer. Cross-reference claims against established medical sources, and don't assume AI synthesis equals verified information.
The study also suggests that health anxiety doesn't necessarily drive better decision-making about information sources. Whether you're mildly concerned or quite worried about symptoms, you're likely to prefer familiar, transparent sources. That's probably wise: high stress can cloud judgment, so using the same trusted sources consistently may be better than switching approaches based on worry level.
For colorectal cancer specifically, if symptoms concern you, don't rely on any online tool as a substitute for medical evaluation. Use information seeking to decide whether to contact your doctor, not to diagnose or dismiss concerns.
| Detail | Value |
|---|---|
| Study type | Randomized controlled trial (2×2 factorial design) |
| Sample size | 762 adults |
| Location | Texas, USA |
| Technology compared | Google search vs. ChatGPT |
| Conditions tested | High-concern (serious symptoms) vs. low-concern (minor symptoms) scenarios |
| Primary measurements | Technology Acceptance Model II constructs, multidimensional trust scales, threat perception measures |
| Primary finding | Google rated significantly higher on usefulness, ease of use, and trustworthiness across all conditions |
| Strongest predictor | Trust as most influential factor for behavioral intention to use |
| Interaction effects | None found between technology type and concern level |
| Evidence quality | A tier (randomized design, validated measures, adequate sample size) |
| Journal | Cancer Control: Journal of the Moffitt Cancer Center |
| PubMed ID | 42171538 |
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3), 319-340.
Kamal, S. A., Shafiq, M., & Kakria, P. (2020). Investigating acceptance of telemedicine services through an extended technology acceptance model (TAM). Technology in Society, 60, 101212.
Witte, K. (1992). Putting the fear back into fear appeals: The extended parallel process model. Communication Monographs, 59(4), 329-349.
Study: Trust and Technology Acceptance: Comparing Traditional Search Engines and Artificial Intelligence for Colorectal Cancer Information Seeking. Published in Cancer Control: Journal of the Moffitt Cancer Center.
ProtocolEngine provides general health information based on published research. This is not medical advice. Consult a healthcare professional before starting any supplement or health protocol.