The BaseResearcharXiv
ScamWorld: An Interactive Environment for Evaluating Web Agents Against Online Fraud
According to the abstract, Yutao Wu, Xiao Liu, Ming Wen and three co-authors built ScamWorld, an environment that reconstructs reported scam websites, and tested web agents powered by six frontier language models. They report baseline deception rates of 11% to 35%, with payments proceeding more readily than credential requests.
Why it matters
The same thing, for different readers.
What it means for you
Researchers report that AI agents shopping or paying for a person can be fooled by fake stores and scams, and proceed with payments more readily than with handing over credentials.
For researchers
Wu, Liu, Wen and co-authors offer a repeatable test bed built from documented scam sites, which are otherwise short-lived and hard to evaluate.
How the finding was reached
In our words, from the source. Weigh the finding by how it was made.
- What was done
- An interactive environment reconstructing reported scam websites across scenarios, fraud mechanisms and languages, used to test web agents powered by six frontier language models.
- What it found
- Baseline deception rates range from 11% to 35% across models; agents proceed more readily with payments than with requests for credentials or wallet information.
- What it does not show
- Only the abstract excerpt was read, so mitigations the authors test and the models tested are not covered here.
- Status
- A preprint: not yet peer reviewed
The record
- Questions
- Are AI companies putting stronger safeguards into practice? · How is AI changing childhood, health and daily life?
- People
- Yutao Wu (Wrote it) · Xiao Liu (Wrote it) · Ming Wen (Wrote it)
- What we read
- We have read only the description the publisher gives. No quotations until someone reads the source itself.
- Found by
- The Radar, run of November 4, 2026, via
arxiv-ai-safety-query(tier 1: the source’s own channel or an official record) - Cite this
- arXiv, “ScamWorld: An Interactive Environment for Evaluating Web Agents Against Online Fraud”, November 3, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/arxiv-2026-11-03-scamworld-web-agents-fraud