arXiv cs.AI· Alexander Loth, Martin Kappes, Marc-Oliver Pahl·· 14 小时前AI 评分31
验证危机：专家如何看待 GenAI 虚假信息与可复现溯源
The Verification Crisis: Expert Perceptions of GenAI Disinformation and the Case for Reproducible Provenance
AI 导读
一项针对 21 名 AI 研究者、政策制定者和虚假信息专家的纵向专家感知调查发现，深度伪造视频带来即时“冲击”效应，但大规模文本生成在政治领域构成“认知碎片化”与“合成共识”的系统性风险。专家对技术检测工具持怀疑态度，更倾向溯源标准与监管框架，尽管落地存在障碍。研究提出将信息完整性视为基础设施，强调数据溯源与方法可复现性，并呼吁建立标准化 benchmark 与可复现性清单。
正文 · 原文
Abstract:The growth of Generative Artificial Intelligence (GenAI) has shifted disinformation production from manual fabrication to automated, large-scale manipulation. This article presents findings from the first wave of a longitudinal expert perception survey (N=21) involving AI researchers, policymakers, and disinformation specialists. It examines the perceived severity of multimodal threats -- text, image, audio, and video -- and evaluates current mitigation strategies. Results indicate that while deepfake video presents immediate "shock" value, large-scale text generation poses a systemic risk of "epistemic fragmentation" and "synthetic consensus," particularly in the political domain. The survey reveals skepticism about technical detection tools, with experts favoring provenance standards and regulatory frameworks despite implementation barriers. GenAI disinformation research requires reproducible methods. The current challenge is measurement: without standardized benchmarks and reproducibility checklists, tracking or countering synthetic media remains difficult. We propose treating information integrity as an infrastructure with rigor in data provenance and methodological reproducibility.
| Comments: | 9 pages. Author version; the Version of Record appeared in the Companion Proceedings of the ACM Web Conference 2026 (WWW '26 Companion), pp. 980-988 |
| Subjects: | Computers and Society (cs.CY); Artificial Intelligence (cs.AI); Social and Information Networks (cs.SI) |
| ACM classes: | K.4.1; K.4.2; I.2.7; H.1.2 |
| Cite as: | arXiv:2602.02100 [cs.CY] |
| (or arXiv:2602.02100v2 [cs.CY] for this version) | |
| https://doi.org/10.48550/arXiv.2602.02100 arXiv-issued DOI via DataCite |
|
| Journal reference: | Companion Proceedings of the ACM Web Conference 2026 (WWW '26 Companion), pp. 980-988 |
| Related DOI: | https://doi.org/10.1145/3774905.3795484
DOI(s) linking to related resources |
Submission history
From: Alexander Loth [view email]
[v1]
Mon, 2 Feb 2026 13:45:12 UTC (23 KB)
[v2]
Thu, 29 Oct 2026 11:35:05 UTC (26 KB)
来源：arXiv cs.AI · arxiv.org