Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2511.02367 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917058164293632 |
|---|---|
| author | Zhang, Shuning Li, Zhaoxin Wen, Changxi Ma, Ying Li, Simin Zhang, Gengrui Zhang, Ziyi Meng, Yibo Zhao, Hantao Yi, Xin Li, Hewu |
| author_facet | Zhang, Shuning Li, Zhaoxin Wen, Changxi Ma, Ying Li, Simin Zhang, Gengrui Zhang, Ziyi Meng, Yibo Zhao, Hantao Yi, Xin Li, Hewu |
| contents | The proliferation of Vision-Language Models (VLMs) introduces profound privacy risks from personal videos. This paper addresses the critical yet unexplored inferential privacy threat, the risk of inferring sensitive personal attributes over the data. To address this gap, we crowdsourced a dataset of 508 everyday personal videos from 58 individuals. We then conducted a benchmark study evaluating VLM inference capabilities against human performance. Our findings reveal three critical insights: (1) VLMs possess superhuman inferential capabilities, significantly outperforming human evaluators, leveraging a shift from object recognition to behavioral inference from temporal streams. (2) Inferential risk is strongly correlated with factors such as video characteristics and prompting strategies. (3) VLM-driven explanation towards the inference is unreliable, as we revealed a disconnect between the model-generated explanations and evidential impact, identifying ubiquitous objects as misleading confounders. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_02367 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | The Pervasive Blind Spot: Benchmarking VLM Inference Risks on Everyday Personal Videos Zhang, Shuning Li, Zhaoxin Wen, Changxi Ma, Ying Li, Simin Zhang, Gengrui Zhang, Ziyi Meng, Yibo Zhao, Hantao Yi, Xin Li, Hewu Human-Computer Interaction The proliferation of Vision-Language Models (VLMs) introduces profound privacy risks from personal videos. This paper addresses the critical yet unexplored inferential privacy threat, the risk of inferring sensitive personal attributes over the data. To address this gap, we crowdsourced a dataset of 508 everyday personal videos from 58 individuals. We then conducted a benchmark study evaluating VLM inference capabilities against human performance. Our findings reveal three critical insights: (1) VLMs possess superhuman inferential capabilities, significantly outperforming human evaluators, leveraging a shift from object recognition to behavioral inference from temporal streams. (2) Inferential risk is strongly correlated with factors such as video characteristics and prompting strategies. (3) VLM-driven explanation towards the inference is unreliable, as we revealed a disconnect between the model-generated explanations and evidential impact, identifying ubiquitous objects as misleading confounders. |
| title | The Pervasive Blind Spot: Benchmarking VLM Inference Risks on Everyday Personal Videos |
| topic | Human-Computer Interaction |
| url | https://arxiv.org/abs/2511.02367 |