Direkt zum Inhalt


Aktuelles: Full Paper auf der CIKM 2025

von Pia Donabauer und David Elsweiler

14. November 2025, von Melanie A. Kilian

  • Informatik und Data Science
  • Forschung
  • Publikation

Can a simple warning change how we respond to AI answer boxes in search engines? 🧐💡

The paper “Exploring the Impact of Warnings on User Perception towards AI-Generated Content in Search Results” by Pia Donabauer and David Elsweiler takes a close look at this question. 🔍 Their study investigates whether a lightweight disclosure banner can help users engage more critically with AI-generated answers.


⚠️ Why does this matter?

As AI answer boxes become standard in search engines, their influence grows — even when the content may not always be reliable.


📊 What the study found:

A simple warning did not shift overall opinions, but it made users 83% more likely to challenge the AI’s stance.

Agreement is a strong anchor: When users initially agreed with the AI, opinion change dropped by 85%, and exploration of opposing results fell significantly.

When the AI disagreed with users, 41% of the participants' post-task explanations drifted linguistically toward the chatbot’s wording (vs. 14% when they already agreed).


💡 Takeaway:

Transparency banners can encourage users to adopt a more critical stance toward AI output — but the are not a complete solution. Additional bias-mitigation strategies are needed to support balanced information seeking in AI-augmented search. 🛠️🤖

 

📍 The study was presented at ACM CIKM 2025, one of the major international forums for Information Retrieval and Knowledge Management, held from November 10–14 in Seoul, Republic of Korea.


Finally, this work highlights just how impressive our students are — and how committed we are to supporting them in publishing their research:
Pia Donabauer conducted this study as part of an undergraduate seminar paper, and David Elsweiler supervised the project and guided it all the way to publication. A fantastic achievement by both! 🎉👏


#AIGeneratedContent #SearchEngines
#BiasMitigation #Transparency
#InformationRetrieval #HumanComputerInteraction
#Research #AI #AIResearch
#EmpiricalStudy
#ACM #ConferenceOnInformationAndKnowledgeManagement #CIKM #CIKM2025
#ResearchSuccess #StudentSuccess #ResearchPaperAccepted
#InformationScienceRegensburg #StayInformed

 


Informationen/Kontakt

Zu David Elsweiler

nach oben