Airbnb is at war with the party people. For several years, the platform that connects owners and vacationers has been multiplying initiatives to put an end to rentals intended to organize festive evenings without the consent of the host. The latest example: the integration into the platform of a tool based on artificial intelligence (AI) capable of detecting and eliminating potential troublemakers, the BBC reported on October 25.
To do this, the AI considers a series of criteria linked to the user. Among them, the date of creation of his Airbnb account, and if he is trying to rent a property in the city where he is domiciled. The system also questions the length of stay and examines whether the visit is planned during a holiday period.
“It’s probably a party if a person books a room around New Year’s Eve, just for one night, and lives in the same city as the host,” Naba Banerjee, Airbnb’s head of security and trust, analyzes for the BBC.
In this case, the AI prevents the reservation or guides the customer to the website of one of Airbnb’s partner hotels if it deems the risk of booking a party to be too high.
Risk of discrimination
Interviewed by the BBC, Edward McFowland III, assistant professor of technology and operations management at Harvard Business School, believes that using AI to filter out potentially problematic customers “can help reduce friction on both sides, both for the company as for the consumer”.
However, despite the laudable goal of this technology, it is also likely to prevent certain categories of people from renting on Airbnb and therefore lead to discrimination.
The professor points out that no matter how sophisticated an AI is, it can still make mistakes and exclude people who simply want to rent an apartment for New Year’s Eve, without having the intention of organizing a party.
“Artificial intelligence technology is still very difficult to develop,” Edward McFowland III told the BBC. Airbnb, for its part, believes that the tool will improve over time thanks to its machine learning and the extensive data collected by the platform.
Source: BFM TV
