Corrigendum to “revealing the influence of AI and its interfaces on job candidates' honest and deceptive impression management in asynchronous video interviews” [Technol. Forecast. Soc. change, 198 (2024), 123,011] (Technological Forecasting & Social Change (2024) 198, (S0040162523006960), (10.1016/j.techfore.2023.123011))

To ensure research transparency and methodological clarity, we provide additional details on our data sources, research design, and statistical reporting. Given the involvement of real job candidates in an AI-AVI hiring process, all data were collected under strict confidentiality agreements, with only de-identified self-reported survey responses used for analysis. This study was conducted as part of a two-year funded research project (Project No. 110–2511-H-003-044-MY2), investigating the effects of AI-assisted Asynchronous Video Interviews (AI-AVI) on candidate impression management behaviors and interview anxiety in employment screening contexts. The research design included five mutually exclusive experimental conditions, with each participant randomly assigned to one of the following groups: Group 1: Non-AI-AVI (Control Group). Group 2: AI-AVI without interface manipulation (Baseline). Group 3: AI-AVI with a tangible interface (Avatar). Group 4: AI-AVI with an immediate interface (Chatbot). Group 5: AI-AVI with a transparent interface (Text-based AI explanation). The study was conducted in collaboration with a Professional Employer Organization (PEO), which facilitated all aspects of participant recruitment and experimental administration. The research team did not have direct access to participants' personal information, receiving only anonymized datasets consistent with predefined specifications. All data were securely transferred by the PEO in compliance with ethical guidelines and data protection standards. Given the sensitive nature of Impression Management (IM) behaviors and interview anxiety, the survey measuring these constructs was independently conducted by the research team. To maintain participant anonymity and ensure unbiased responses, email invitations distributed by the PEO directed candidates to the research team's independent online questionnaire system. Participants provided informed consent electronically within the survey system, confirming that responses would remain confidential and not shared with third parties, including the PEO. All data were fully anonymized prior to analysis. No interview transcripts were recorded or analyzed. Data processing was conducted to ensure completeness and accuracy, adhering to ethical standards and prioritizing participant privacy. In the original submission, the correlation coefficients among the four manipulated AI-AVI interface conditions were omitted from Table 3 due to the categorical and mutually exclusive nature of the experimental groups. Pearson correlations typically apply to continuous variables; here, conditions were dummy-coded (0/1), rendering direct correlation calculations statistically less meaningful (O'Grady & Medoff, 1988). O'Grady, K. E., & Medoff, D. R. (1988). Categorical Variables in Multiple Regression: Some Cautions. Multivariate Behavioral Research, 23(2), 243–260. https://doi.org/10.1207/s15327906mbr2302_7 As described in the paper: “The intercorrelations among AI and the other variables were computed based on observations from the AI-AVI without the three interfaces and the non-AI-AVI group (Analogy 1). In contrast, the intercorrelations among tangibility, immediacy, transparency, and the other variables were computed solely based on observations from the AI-AVI groups (Analogies 2, 3, and 4).” Therefore, the correlation coefficients were computed from 130 participants assigned to AI-based AVIs (Groups 2–5), excluding the non-AI-AVI condition (Group 1). The dummy coding for manipulated variables was: AI: AI interface presence (1) vs. none (0). Tangibility: Avatar presence (1) vs. none (0). Immediacy: Chatbot presence (1) vs. none (0). Transparency: Text-based AI explanation (1) vs. none (0). These groups were mutually exclusive; therefore, correlation coefficients among these dummy-coded variables should be interpreted with caution. Nonetheless, for transparency, we provide these coefficients in the revised Table 3, with the caveat regarding their interpretative limitations. This paper and Suen & Hung (2023) originated from the same two-year funded research project (Project No. 110–2511-H-003-044-MY2), conducted by the same research team and employing identical experimental interfaces and participant recruitment processes. However, the two studies differed substantively in their research questions, theoretical frameworks, hypotheses, dependent variables, and analytical approaches. Suen & Hung (2023), completed in 2022 and published in 2023, focused on candidates' cognitive and affective trust perceptions based on Trust in Automation Theory and Social Interface Theory. In contrast, the current study, conducted from 2022 to 2023 and published in 2024, specifically examined candidates' honest and deceptive impression management behaviors through Impression Management Theory and the Faking Likelihood Model. Therefore, although both studies share the same overarching research context and interface classification (as proposed by Suen & Hung, 2023), each uniquely contributes to theoretical and practical understandings of AI-driven hiring without redundancy. Key Differences Between This Paper and Suen & Hung (2023): [Table presented] Specifically, Suen & Hung (2023) focused on candidates' cognitive and affective trust perceptions in AI-based interview systems, contributing to the literature on trust in automation and social interface theory. In contrast, this study examined candidates' impression management behaviors under various AI interface conditions, providing novel insights that align with impression management and employment faking research. By explicitly adopting the interface classification framework proposed by Suen & Hung (2023), this study ensured conceptual coherence while advancing both theoretical and practical knowledge in AI-driven hiring contexts. Collectively, these two complementary studies provided comprehensive yet distinct contributions to understanding AI's role in asynchronous video interviews, without concerns regarding redundant or fragmented publication.