Age EstimationAdversarialBenchmark

Can a Teenager Fool an AI? Evaluating Low-Cost Cosmetic Attacks on Age Estimation Systems

Xingyu Shen, Tommy Duong, Xiaodong An, Zengqi Zhao, Zebang Hu, Haoyu Hu, Ziyou Wang, Finn Guo, Simiao Ren

Read on arXiv

How to cite

@article{shen2026can,
  title={Can a Teenager Fool an AI? Evaluating Low-Cost Cosmetic Attacks on Age Estimation Systems},
  author={Shen, Xingyu and Duong, Tommy and An, Xiaodong and Zhao, Zengqi and Hu, Zebang and Hu, Haoyu and Wang, Ziyou and Guo, Finn and Ren, Simiao},
  journal={arXiv preprint arXiv:2602.19539},
  year={2026}
}

Associated datasets

Adversarial age estimation attack dataset

Faces with low-cost cosmetic adversarial perturbations designed to defeat age estimation systems — used in 'Can a Teenager Fool an AI?'.

More from Age Estimation

Try it in production

Scam AI's Eva-v1 model applies this research at production scale — 98.2% accuracy, under 4 seconds, 200 free detections per month.

AI Image & Video DetectionStart free