paper: <a href=\"https://arxiv.org/pdf/2606.03348\" rel=\"nofollow\">https://arxiv.org/pdf/2606.03348</a><br>github: <a href=\"https://github.com/thu-coai/Syncred-Bench\" rel=\"nofollow\">https://github.com/thu-coai/Syncred-Bench</a><br>data: <a href=\"https://huggingface.co/datasets/thu-coai/Syncred-Bench\">https://huggingface.co/datasets/thu-coai/Syncred-Bench</a></p>\n","updatedAt":"2026-06-04T00:37:40.436Z","author":{"_id":"65d859a3661492b25c46a117","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/65d859a3661492b25c46a117/Yui5RSHsWltBF3s3X4NI2.jpeg","fullname":"Junxiao Yang","name":"yangjunxiao2021","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":5,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.6432642936706543},"editors":["yangjunxiao2021"],"editorAvatarUrls":["https://cdn-avatars.huggingface.co/v1/production/uploads/65d859a3661492b25c46a117/Yui5RSHsWltBF3s3X4NI2.jpeg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2606.03348","authors":[{"_id":"6a202b7615100c5272a842d1","name":"Junxiao Yang","hidden":false},{"_id":"6a202b7615100c5272a842d2","name":"Minghao Zhang","hidden":false},{"_id":"6a202b7615100c5272a842d3","name":"Xiaoce Wang","hidden":false},{"_id":"6a202b7615100c5272a842d4","name":"Haoran Liu","hidden":false},{"_id":"6a202b7615100c5272a842d5","name":"Shiyao Cui","hidden":false},{"_id":"6a202b7615100c5272a842d6","name":"Hongning Wang","hidden":false},{"_id":"6a202b7615100c5272a842d7","name":"Minlie Huang","hidden":false}],"publishedAt":"2026-06-02T00:00:00.000Z","submittedOnDailyAt":"2026-06-03T00:00:00.000Z","title":"SynCred-Bench: Benchmarking Synthetic Credibility in AI-Generated Visual Misinformation","submittedOnDailyBy":{"_id":"65d859a3661492b25c46a117","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/65d859a3661492b25c46a117/Yui5RSHsWltBF3s3X4NI2.jpeg","isPro":false,"fullname":"Junxiao Yang","user":"yangjunxiao2021","type":"user","name":"yangjunxiao2021"},"summary":"Recent generative models can now produce visual artifacts with realistic embedded text and layouts, creating a new misinformation threat: synthetic credibility. We introduce SYNCRED-Bench, a benchmark of 600 AI-generated misinformation images balanced across six credible-form categories and seven fine-grained circulation styles, together with FP450, a real-image negative set for measuring false positives. Extensive evaluation shows that existing systems remain unreliable: under a 5% false-positive-rate constraint, 15 MLLMs achieve only 10.5% true positive rate (TPR), open-source AIGC detectors achieve less than 5%, and commercial APIs reach 57.6%. Human annotators also struggled to identify synthetic credibility, reaching only 63% TPR. These findings establish synthetic credibility as a severe and underexplored visual misinformation challenge, and provide a benchmark for developing detectors that reason beyond superficial credibility cues.","upvotes":1,"discussionId":"6a202b7615100c5272a842dd","ai_summary":"AI-generated images with realistic text and layouts pose a significant misinformation threat requiring new detection benchmarks and methods beyond surface-level credibility assessment.","ai_keywords":["generative models","misinformation","synthetic credibility","SYNCRED-Bench","FP450","MLLMs","AIGC detectors","commercial APIs","false-positive-rate","true positive rate"],"ai_summary_model":"Qwen/Qwen2.5-Coder-32B-Instruct"},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"65d859a3661492b25c46a117","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/65d859a3661492b25c46a117/Yui5RSHsWltBF3s3X4NI2.jpeg","isPro":false,"fullname":"Junxiao Yang","user":"yangjunxiao2021","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0}">
SynCred-Bench: Benchmarking Synthetic Credibility in AI-Generated Visual Misinformation
Abstract
AI-generated images with realistic text and layouts pose a significant misinformation threat requiring new detection benchmarks and methods beyond surface-level credibility assessment.
Recent generative models can now produce visual artifacts with realistic embedded text and layouts, creating a new misinformation threat: synthetic credibility. We introduce SYNCRED-Bench, a benchmark of 600 AI-generated misinformation images balanced across six credible-form categories and seven fine-grained circulation styles, together with FP450, a real-image negative set for measuring false positives. Extensive evaluation shows that existing systems remain unreliable: under a 5% false-positive-rate constraint, 15 MLLMs achieve only 10.5% true positive rate (TPR), open-source AIGC detectors achieve less than 5%, and commercial APIs reach 57.6%. Human annotators also struggled to identify synthetic credibility, reaching only 63% TPR. These findings establish synthetic credibility as a severe and underexplored visual misinformation challenge, and provide a benchmark for developing detectors that reason beyond superficial credibility cues.
Community
Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images
Cite arxiv.org/abs/2606.03348 in a model README.md to link it from this page.
Cite arxiv.org/abs/2606.03348 in a dataset README.md to link it from this page.
Cite arxiv.org/abs/2606.03348 in a Space README.md to link it from this page.
Discussion (0)
Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.
Sign in →No comments yet. Sign in and be the first to say something.