Logo of Homebase Zukunft Podcast
Fake shops cause major economic damage; a dark field study estimates that 320,000 consumers in Austria are directly affected and puts the amount of damage at 16 million euros. Against this backdrop, it is clear that appropriate measures must be taken to protect online customers from this form of fraud.
The AIT has developed the Fake-Shop Detector (FSD) (https://fakeshop.at) for this purpose as part of projects funded by the BMF in the KIRAS security research funding programme, among others. The FSD is a free service that warns of fraudulent online retailers directly when surfing in the browser and thus proactively and effectively protects consumers when shopping online. The Detector uses over 10,000 known fake shops in German-speaking countries. In addition, unknown websites are analysed by the Detector in real time using artificial intelligence and evaluated in terms of their similarity to already known fraudulent providers.
The decisive factor in calculating the threat potential here is not individual features, but a large number of features in combination. Over 21,500 features are included in the real-time risk assessment by artificial intelligence. Since the end of July 2020, 863,114 domains have been analysed by the AI of the Fake Shop Detector, achieving a model accuracy of over 91% in practical use.
Another key feature of the Fake Shop Detector is the trustworthy AI, which is ensured by manual quality assurance carried out on an ongoing basis by experts from Watchlist Internet. This is particularly important as the FSD is a security-relevant service aimed at consumers, who should be protected during online shopping.
Solvion and Peter Hössl are the founders of the podcast and have of course scrutinised how exactly AI affects this topic and, in addition to the Fake Shop Detector (https://www.fakeshop.at/) have also given thought to the topic in general: