Airbnb builds software to judge guest/host reliability?

Patent data revealed that Airbnb is trying to build software that predicts users’ credibility and suitability, behavior and personality traits through Internet activity.

This is according to a patent published in the United States in 2019. According to the patent data, the software developed by Airbnb searches websites including social media to make a database of credit information and identity verification, etc. Check if there is a tendency.

In addition, if it finds a fake social network profile, AI will rate the user. If related keywords, photos, or videos such as drugs or alcohol related sites are found, the user score decreases. Likewise, negative comments about online content related to performance result in low evaluation.

This patent searches or finds news that may be related to users through machine learning, and assigns a weight to the severity of violations. Also, posting on a blog or website is considered. It analyzes how users behave with others in reality through such data and information such as social relationships, employment, and education. In this way, both the guest and the host calculate suitability.

Airbnb has not commented in detail, saying that the background check startup that it acquired for the patent is an extension of Trolly. Airbnb says that all its reservations are risk assessed until they are confirmed, and that it is conducting an investigation that detects suspicious movements before problems occur, assessing them through predictive analytics and machine learning. However, there are voices of concern about the use of personal information when performing such tests. Related information can be found here .