Bots and Beyond: Threats to Trust In Online Caregiver Research

Background and Purpose: Internet-based survey methods are widely used in social work research because they are accessible, cost-effective, and able to reach hard-to-reach populations such as informal caregivers. However, the rapid emergence of bots and artificial intelligence (AI)-powered fraudulent respondents poses significant threats to data integrity. While prior research has focused on technical strategies to detect and prevent fraudulent responses, less attention has been paid to how implementing these strategies affects trust, participant burden, and researcher-participant relationships. Centering on trust and relationships, which are core to social work research, this study describes how researchers’ fraud detection and prevention practices influence trust-building in online caregiver research.

Methods:
This study employs a reflexive qualitative case-study approach, drawing on researchers’ positionality and experiences from a federally funded caregiver study that used hybrid recruitment and online data collection. We present four illustrative cases encountered during study recruitment and screening, including 1) bot attacks following social media dissemination, 2) suspicious referrals from established research registries, 3) ambiguous phone-based screening interactions, and 4) the misclassification of legitimate participants as fraudulent. Through reflexive analysis, we described how both fraudulent activities and the prevention strategies implemented influenced researchers’ decision-making, researcher-participant interactions, and researchers’ relationships with participants and the community.

Results:
Findings revealed that AI-enabled fraud disrupts trust and relationships in several ways. First, the uncertainty about participant authenticity led researchers to self-doubt and mistrust with participants during phone and email interactions, making it difficult to rely on their own prior experience to make decisions. Second, the relationships between researchers and participants changed, with researchers assuming a gatekeeping role that positioned participants as potentially deceptive, challenging the core social work values of respect and partnership. Third, trust in recruitment channels (e.g., online research registries and community facing platforms) was weakened, raising concerns about accountability and the reliability of these infrastructures. Fourth, verification processes introduced burden to participants, particularly caregivers with limited time and energy, who may perceive additional screening as burdensome or as signals of distrust. Finally, efforts to prevent fraudulent participation created tensions between ethical inclusion and exclusion: in trying to filter out fraudulent responses, researchers risked excluding legitimate participants, particularly those already underrepresented.

Conclusions and Implications:
Bots and fraudulent respondents are threats to trust and relationships in online caregiver research across multiple stages. While fraud prevention is necessary for research, it may unintentionally weaken relationships with participants and communities, increase burden, and limit access for those most in need of being heard. To address these challenges, efforts to protect data integrity must be balanced with the need to preserve participants’ confidentiality, minimize burden, and ensure equitable access. Transparent communication about the verification processes, attention to participants’ experience, and the integration of human-centered recruitment strategies are essential to rebuilding trust. At the institutional level, there is a need for developing shared guidelines, collective responses, and technological solutions to ethical fraud prevention that centers on researcher-participant relationships. Future research should further explore how to design trustworthy and inclusive digital research practices in the context of the evolving technological landscape.