Fake-Shop Detector

Functionality

The Fake-Shop Detector checks every online shop in two steps: First, it searches a database of reputable and fraudulent online shops, which is curated by experts. If an online shop is unknown, a real-time analysis is performed using artificial intelligence (AI).

During the automatic analysis of websites, attention is paid to the quality and traceability of the AI. It never learns unsupervised, but is subject to constant review and quality control by the experts at Watchlist Internet. Since the end of July 2020, more than 1 million shop evaluations have already been carried out using the Fake-Shop Detector.

The Fake Shop Detector knows more than 10,000 fraudulent online shops and more than 26,000 trustworthy online shops in the DACH region (Germany, Austria, Switzerland). Currently, over 8,500 users take advantage of the offer and use the plugin. The system receives 400,000 API requests per day.

Curated Database

In a first step, the Fake-Shop Detector uses a database to check whether the page you have accessed has already been clearly identified as a fake shop or a reputable online shop. This database includes the following data sources, which have been verified by experts:

• List of fraudulent or problematic online shops from Watchlist Internet
• Shops with seals of approval and certificates (such as the Austrian e-commerce seal of approval)
• Lists of cooperation partners
• Directories of Austrian online shops
• Entries on reputable online shops from previous detector checks

Real-time Analysis using Artificial Intelligence

All unknown websites are evaluated by the Fake-Shop Detector in terms of their similarity to fake shops. The decisive factors are a combination of many characteristics. Over 21,500 characteristics are incorporated into the real-time evaluation by artificial intelligence.

To ensure quality, the artificial intelligence behind it was developed in cooperation with the Internet fraud experts at Watchlist Internet. After all, AI is only as good as the data on which it is based. The data basis for the Fake-Shop Detector are thousands of known fake shops, from which page characteristics were taken. The quality of the risk assessments is also continuously checked by the experts. Past tests show that the risk assessments of the Fake-Shop Detector are very good overall.

Data Protection

The protection of your personal data is very important to us. As a general rule, the Fake-Shop Detector follows the principle of data minimisation as set out in the GDPR. Apart from your IP address and a plugin ID, which is randomly generated when the plugin is installed, no other personal data is recorded when requests are sent to the server. The detector's privacy-by-design approach ensures that a large number of already known risk assessments are answered offline and directly locally from the cache of your own browser. This means that even in the event of a breach of our backend system, it is impossible to create a profile of individual users' surfing behaviour from the log files. The terms of use stipulate that data from the log files will never be merged or passed on to third parties.