Amazon wants to sniff out shady messages between buyers and sellers, a patent filing reveals. Two big reasons for concern: fraud and fee avoidance.
Amazon’s patent cited examples of how unscrupulous buyers might attempt to engage in various types of fraudulent activities:
- submit orders without providing valid payment (e.g., using fraudulent money orders, stolen credit cards, etc.);
- attempt to obtain personal information about sellers;
- attempt to move communications off the marketplace to avoid protections provided by or fees charged by the electronic marketplace.
Amazon also provided examples of messages unscrupulous sellers might send:
- attempt to defraud buyers;
- obtain unauthorized access to sellers’ accounts and then impersonate the sellers.
In the later case, Amazon pointed out that with access to a seller’s account, a fraudster could gain access to contact information for buyers with whom the seller has previously interacted, which the user may then use for various types of fraudulent activities (e.g., to “phish” for other confidential information about the buyers).
“Thus,” Amazon writes, “it would be beneficial to provide techniques to inhibit fraudulent activities related to electronic commerce, including by inhibiting electronic communications that reflect or further such fraudulent activities, as well as to provide other benefits.”
The patent filing describes various techniques for assessing communications to identify those that are suspect. Amazon might then block the communication from reaching its intended recipient, or it could flag the communication as suspect, or it could modify the message to remove inappropriate content.
You can find the patent filing on the USPTO website. Can technology get it right, or is it inevitable that it miss some fraud and misidentify some non-fraudulent messages?