Metraipolis › archive › dc_e09577fac15c

DIBench: Benchmarking Decision Integrity of GUI-based Mobile Agents Under Deceptive Injections — arXiv:2610.06898

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.06898
License (as archived): CC0-1.0
Archived: 2026-10-07T11:17:14.048Z
SHA-256 of the archived text: c6c431ac7e0b62b8ace05df0052bbb31c77c44a2ac6a903851e16953ebf3fd91
Stable URL of this page: https://metraipolis.com/documento/dc_e09577fac15c

How to cite

DIBench: Benchmarking Decision Integrity of GUI-based Mobile Agents Under Deceptive Injections — arXiv:2610.06898 — arXiv (en), https://arxiv.org/abs/2610.06898 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e09577fac15c (Metraipolis archive dc_e09577fac15c, sha256 c6c431ac7e0b62b8ace05df0052bbb31c77c44a2ac6a903851e16953ebf3fd91)

BibTeX

@misc{dce09577fac15c,
  title = {DIBench: Benchmarking Decision Integrity of GUI-based Mobile Agents Under Deceptive Injections — arXiv:2610.06898},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.06898},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e09577fac15c; sha256 c6c431ac7e0b62b8ace05df0052bbb31c77c44a2ac6a903851e16953ebf3fd91},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.06898. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

As GUI-based mobile agents rapidly progress, rigorous safety evaluation of their autonomous decision-making in realistic app interfaces becomes increasingly critical. Existing benchmarks mainly focus on execution-level anomalies using task success or hijack rates, but fail to capture the in-task goal deviation risk in multi-candidate selection tasks, where the decision may be steered toward an attacker-specified target, even in violation of instruction-implied constraints (e.g., cheapest/highest-rated), without any overt execution anomalies. We present DIBench, a decision integrity benchmark for measuring this risk in mobile agents. DIBench covers 7 commercial and 3 simulated apps with 5 task types. Under a threat model restricted to non-privileged UI content, we construct 8 deceptive injection probe instantiations that can steer critical selections without overt anomalies. The benchmark includes 1,000 clean and 36,672 injected instances, with a unified protocol and integrity metrics for comparison. Experiments spanning 4 agent frameworks and 7 base models show that completion-based evaluation can overestimate agent trustworthiness and miss decision-integrity risks: deceptive injections steer selections and shift early action policies, inflating completion rates and creating a misleading illusion of safety. Common defenses, including detection, image preprocessing, and prompt reminders, yield inconsistent integrity gains. Overall, DIBench provides a unified, reproducible benchmark to quantify the risk of in-task goal deviation in mobile agents and enable comparable evaluations of safety defenses.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.MA, nuovi lavori del giorno).
Categorie arXiv: Cryptography and Security (cs.CR) ; Multiagent Systems (cs.MA).
Pagina del lavoro: https://arxiv.org/abs/2610.06898 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-07T11:15:09.465Z dalla pagina https://arxiv.org/list/cs.MA/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).