Research

Paper

AI LLM March 11, 2026

Naïve Exposure of Generative AI Capabilities Undermines Deepfake Detection

Authors

Sunpill Kim, Chanwoo Hwang, Minsu Kim, Jae Hong Seo

Abstract

Generative AI systems increasingly expose powerful reasoning and image refinement capabilities through user-facing chatbot interfaces. In this work, we show that the naïve exposure of such capabilities fundamentally undermines modern deepfake detectors. Rather than proposing a new image manipulation technique, we study a realistic and already-deployed usage scenario in which an adversary uses only benign, policy-compliant prompts and commercial generative AI systems. We demonstrate that state-of-the-art deepfake detection methods fail under semantic-preserving image refinement. Specifically, we show that generative AI systems articulate explicit authenticity criteria and inadvertently externalize them through unrestricted reasoning, enabling their direct reuse as refinement objectives. As a result, refined images simultaneously evade detection, preserve identity as verified by commercial face recognition APIs, and exhibit substantially higher perceptual quality. Importantly, we find that widely accessible commercial chatbot services pose a significantly greater security risk than open-source models, as their superior realism, semantic controllability, and low-barrier interfaces enable effective evasion by non-expert users. Our findings reveal a structural mismatch between the threat models assumed by current detection frameworks and the actual capabilities of real-world generative AI. While detection baselines are largely shaped by prior benchmarks, deployed systems expose unrestricted authenticity reasoning and refinement despite stringent safety controls in other domains.

Metadata

arXiv ID: 2603.10504
Provider: ARXIV
Primary Category: cs.CR
Published: 2026-03-11
Fetched: 2026-03-12 04:21

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.10504v1</id>\n    <title>Naïve Exposure of Generative AI Capabilities Undermines Deepfake Detection</title>\n    <updated>2026-03-11T07:58:38Z</updated>\n    <link href='https://arxiv.org/abs/2603.10504v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.10504v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Generative AI systems increasingly expose powerful reasoning and image refinement capabilities through user-facing chatbot interfaces. In this work, we show that the naïve exposure of such capabilities fundamentally undermines modern deepfake detectors. Rather than proposing a new image manipulation technique, we study a realistic and already-deployed usage scenario in which an adversary uses only benign, policy-compliant prompts and commercial generative AI systems. We demonstrate that state-of-the-art deepfake detection methods fail under semantic-preserving image refinement. Specifically, we show that generative AI systems articulate explicit authenticity criteria and inadvertently externalize them through unrestricted reasoning, enabling their direct reuse as refinement objectives. As a result, refined images simultaneously evade detection, preserve identity as verified by commercial face recognition APIs, and exhibit substantially higher perceptual quality. Importantly, we find that widely accessible commercial chatbot services pose a significantly greater security risk than open-source models, as their superior realism, semantic controllability, and low-barrier interfaces enable effective evasion by non-expert users. Our findings reveal a structural mismatch between the threat models assumed by current detection frameworks and the actual capabilities of real-world generative AI. While detection baselines are largely shaped by prior benchmarks, deployed systems expose unrestricted authenticity reasoning and refinement despite stringent safety controls in other domains.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CR'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.AI'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CV'/>\n    <published>2026-03-11T07:58:38Z</published>\n    <arxiv:primary_category term='cs.CR'/>\n    <author>\n      <name>Sunpill Kim</name>\n    </author>\n    <author>\n      <name>Chanwoo Hwang</name>\n    </author>\n    <author>\n      <name>Minsu Kim</name>\n    </author>\n    <author>\n      <name>Jae Hong Seo</name>\n    </author>\n  </entry>"
}