Haeseung Seo

dblp:244/0106 · DBLP profile ↗
← Back
7ranked-venue papers in the field
2as first author
6since 2021 · last 2025
0000-0003-0393-6691ORCID · corroborated

Domains — venue-derived; a paper can count in several

Information Retrieval & Web Search · 6 (2 first)Data Mining & Knowledge Discovery · 1
YearPublicationVenuePosition
2025 A Study of Training Strategies on Enhancing Human Detection of AI-Synthesized Faces
abstract
Artificial intelligence (AI) synthesized faces—so called deepfake images—have been increasingly used for malicious intent and have resulted in prominently adverse impact. Because online users must contend with discerning fake from real, great emphasis has been placed on enhancing human detection of deepfake images. We conducted an online human-subject study (N=237), investigating the effect of three training strategies (explicit training with visible artifacts in synthetic faces, implicit training with experiencing the generation of synthetic faces using real human faces, and a combination of both artifact and generation) on participants’ detection of synthetic faces generated by the state-of-the-art StyleGAN techniques. Comparing participants’ deepfake detection across three phases (baseline in phase 1 without any training, phase 2 after one training session, and phase 3 after the other training session), we found that all training strategies effectively enhanced participants’ detection of AI-synthesized faces and their decision confidence. We also explored factors that impact participants’ learning and decision-making of deepfake detection. Responses to the open-ended question revealed that participants developed generalized strategies and utilized artifacts beyond the training. Our quantitative and qualitative results provide nuanced insights into the promises and limitations of the training strategies. In addition to advancing theoretical understanding of human training in the context of deepfake image detection, our study findings hold practical implications for interface design.
Ester Chen, Haeseung Seo, Margie Ruffin, Dongwon Lee 0001, Gang Wang 0011, Aiping Xiong
ICWSM2
2025 Partisan Fact-Checkers' Warnings Can Effectively Correct Individuals' Misbeliefs About Political Misinformation
abstract
Political misinformation, particularly harmful when it aligns with individuals' preexisting beliefs and political ideologies, has become widespread on social media platforms. In response, platforms like Facebook and X introduced warning messages leveraging fact-checking results from third-party fact-checkers to alert users against false content. However, concerns persist about the effectiveness of these fact-checks, especially when fact-checkers are perceived as politically biased. To address these concerns, this study presents findings from an online human-subject experiment (N=216) investigating how the political stances of fact-checkers influence their effectiveness in correcting misbeliefs about political misinformation. Our findings demonstrate that partisan fact-checkers can decrease the perceived accuracy of political misinformation and correct misbeliefs without triggering backfire effects. This correction is even more pronounced when the misinformation aligns with individuals' political ideologies. Notably, while previous research suggests that fact-checking warnings are less effective for conservatives than liberals, our results suggest that explicitly labeled partisan fact-checkers, positioned as political counterparts to conservatives, are particularly effective in reducing conservatives' misbeliefs toward pro-liberal misinformation.
Sian Lee, Haeseung Seo, Aiping Xiong, Dongwon Lee 0001
ICWSM2
2024 Does It Matter Who Said It? Exploring the Impact of Deepfake-Enabled Profiles on User Perception towards Disinformation
abstract
Recently, deepfake techniques have been adopted by real-world adversaries to fabricate believable personas (posing as experts or insiders) in disinformation campaigns to promote false narratives and deceive the public. In this paper, we investigate how fake personas influence the user perception of the disinformation shared by such accounts. Using Twitter as an exemplary platform, we conduct a user study (N=417) where participants read tweets of fake news with (and without) the presence of the tweet authors' profiles. Our study examines and compares three types of fake profiles: deepfake profiles, profiles of relevant organizations, and simple bot profiles. Our results highlight the significant impact of deepfake and organization profiles on increasing the perceived information accuracy of and engagement with fake news. Moreover, deepfake profiles are rated as significantly more real than other profile types. Finally, we observe that users may like/reply/share a tweet even though they believe it was inaccurate (e.g., for fun or truth-seeking), which could further disseminate false information. We then discuss the implications of our findings and directions for future research.
Margie Ruffin, Haeseung Seo, Aiping Xiong, Gang Wang 0011
ICWSM2
2024 Reliability Matters: Exploring the Effect of AI Explanations on Misinformation Detection with a Warning
abstract
To mitigate misinformation on social media, platforms such as Facebook have offered warnings to users based on the detection results of AI systems. With the evolution of AI detection systems, efforts have been devoted to applying explainable AI (XAI) to further increase the transparency of AI decision-making. Nevertheless, few factors have been considered to understand the effectiveness of a warning with AI explanations in helping humans detect misinformation. In this study, we report the results of three online human-subject experiments (N = 2,692) investigating the framing effect and the impact of an AI system’s reliability on the effectiveness of AI warning with explanations. Our findings show that the framing effect is effective for participants’ misinformation detection, whereas the AI system’s reliability is critical for humans’ misinformation detection and participants’ trust in the AI system. However, adding the explanations can potentially increase participants’ suspicions on miss errors (i.e., false negatives) in the AI system. Furthermore, more trust is shown in the AI warning without explanations condition. We conclude by discussing the implications of our findings.
Haeseung Seo, Sian Lee, Dongwon Lee 0001, Aiping Xiong
ICWSM1
2023 Associative Inference Can Increase People's Susceptibility to Misinformation
abstract
Associative inference is an adaptive, constructive process of memory that allows people to link related information to make novel connections. We conducted three online human-subjects experiments investigating participants’ susceptibility to associatively inferred misinformation and its interaction with their cognitive ability and how news articles were presented. In each experiment, participants completed recognition and perceived accuracy rating tasks for the snippets of news articles in a tweet format across two phases. At Phase 1, participants viewed real news only. At Phase 2, participants viewed both real and fake news. Critically, we varied whether the fake news at Phase 2 was inferred from (i.e., associative inference), associated with (i.e., association only), or irrelevant to (i.e., control) the corresponding real news pairs at Phase 1. Both recognition and perceived accuracy results showed that participants in the associative inference condition were more susceptible to fake news than those in the other conditions. Furthermore, hashtags embedded within the tweets made the obtained effects evident only for the participants of higher cognitive ability. Our findings reveal that associative inference can be a basis for individuals’ susceptibility to misinformation, especially for those of higher cognitive ability. We conclude by discussing the implications of our results for understanding and mitigating misinformation on social media platforms.
Sian Lee, Haeseung Seo, Dongwon Lee 0001, Aiping Xiong
ICWSM2
2022 If You Have a Reliable Source, Say Something: Effects of Correction Comments on COVID-19 Misinformation
Haeseung Seo, Aiping Xiong, Sian Lee, Dongwon Lee 0001
ICWSM1
2020 DETERRENT: Knowledge Guided Graph Attention Network for Detecting Healthcare Misinformation
abstract
To provide accurate and explainable misinformation detection, it is often useful to take an auxiliary source (e.g., social context and knowledge base) into consideration. Existing methods use social contexts such as users' engagements as complementary information to improve detection performance and derive explanations. However, due to the lack of sufficient professional knowledge, users seldom respond to healthcare information, which makes these methods less applicable. In this work, to address these shortcomings, we propose a novel knowledge guided graph attention network for detecting health misinformation better. Our proposal, named as DETERRENT, leverages on the additional information from medical knowledge graph by propagating information along with the network, incorporates a Medical Knowledge Graph and an Article-Entity Bipartite Graph, and propagates the node embeddings through Knowledge Paths. In addition, an attention mechanism is applied to calculate the importance of entities to each article, and the knowledge guided article embeddings are used for misinformation detection. DETERRENT addresses the limitation on social contexts in the healthcare domain and is capable of providing useful explanations for the results of detection. Empirical validation using two real-world datasets demonstrated the effectiveness of DETERRENT. Comparing with the best results of eight competing methods, in terms of F1 Score, DETERRENT outperforms all methods by at least 4.78% on the diabetes dataset and 12.79% on cancer dataset. We release the source code of DETERRENT at: https://github.com/cuilimeng/DETERRENT.
Limeng Cui, Haeseung Seo, Maryam Tabar, Fenglong Ma, Suhang Wang, Dongwon Lee 0001
KDD2