Airbnb relies on artificial intelligence in the fight against fraudsters

The popular hosting platform Airbnb is struggling with a plague of fictitious offers. So far, 59 thousand false advertisements have been removed from it, and another 157 thousand have been blocked before publication.

Airbnb relies on artificial intelligence in the fight against fraudsters
00:00 00:00

Summary

  • Airbnb users have raised concerns about the authenticity of premises and lack of price transparency, with landlords often offsetting low rental costs with high additional fees.
  • CEO Brian Chesky acknowledges these issues, noting that Airbnb's reputation and trustworthiness are at stake as the platform becomes a haven for scammers.
  • Airbnb plans to use artificial intelligence (AI) to address these problems. The company has been leveraging machine learning and data analysis for years to optimize guest and host experiences.
  • AI is used in various ways on the platform, including classifying photos, adjusting prices based on the season, translating offer descriptions and reviews, preventing disruptive gatherings, and detecting unusual login activity and malware.
  • The company is now planning to use AI to verify fake offers in the United States, the United Kingdom, Canada, France, and Australia. Hosts will be asked to enter their property and open the Airbnb app, with GPS verifying their location and AI comparing current photos with those in the offer.
  • Properties that pass the test will receive a "verified" icon on their offer from February. This is part of a series of major changes, with offer verification in 30 more countries set to begin next year.

Expensive and risky

In a survey conducted among Airbnb users, tenants complain that the company does not verify the authenticity of premises and does not guarantee price transparency. Landlords regularly tempt with low rental costs, making up for it with, for example, a high cleaning fee.

The company's CEO, Brian Chesky, confirms that he has received many opinions from users, according to which the proposed service has become unprofitable and has become a convenient platform for scammers. Meanwhile, Airbnb is forced to bear the costs of refunds and rebookings. And this threatens the loss of trust and reputation.

– If someone, booking accommodation, is not sure that they will rent an existing and satisfactory place on Airbnb, they will choose a hotel – Chesky comments.

AI remedy for everything

The company sees the solution to its problems in neural networks. Airbnb has been using machine learning and data analysis to drive its growth for years. Currently, there are over 7 million rental offers in over 100,000 cities around the world on the platform, and a large part of this growth results from the use of AI and machine learning in optimizing guest and host experiences.

AI classifies photos for Airbnb and handles the Smart Pricing feature, which automatically adjusts to the price based on the season. A machine learning technology called Translation Engine translates offer descriptions, reviews, and messages between guests. The Anti-Party AI system solves the problem of disruptive gatherings, and another model analyzes unusual login activity and looks for malware. 

Now the company is rolling out more heavy artillery and plans to use artificial intelligence to verify fake offers in the five most popular destinations, i.e. in the United States, the United Kingdom, Canada, France, and Australia. Checked hosts will be asked to enter the property and open the Airbnb app. GPS will verify if they are at the correct address, and artificial intelligence will compare current photos with those in the offer. Each property that passes the test positively will receive a "verified" icon on its offer from February.

As the company announces, this is just the beginning of major changes in the service. Verification of offers in the next thirty countries will start at the beginning of next year.