Airbnb activates "Defense" to prevent unauthorized New Year parties

Airbnb has enabled AI-based tools to help identify and prevent unauthorized and disruptive New Year’s Eve parties hosted through the platform.

The property booking platform said it has introduced measures that use machine learning to identify and block potentially risky bookings of entire home listings.

The technology looks at signals such as length of trip, type of listing, distance from the guest’s location and when the reservation is made to determine whether it might be more risky.

Airbnb is committed to supporting hosts and the communities where they live, and we hope these defenses will allow guests, hosts and neighbors to celebrate the holidays with more confidence.

Amanda Caples, Airbnb

If flagged as higher risk, guests will either be blocked from booking or redirected to alternative locations on the platform.

The company began using the restrictions in 2020 and says it has seen a significant drop in the rate of companies reporting to the company since then.

Airbnb said nearly 74,000 people were banned from booking last year alone, including around 7,800 in the UK.

This included around 2,900 bookings in London and 470 in Manchester.

Amanda Caples, general manager of Airbnb for the UK and Northern Europe, said: “While problems are rare, we want to do our part to help reduce the risk of unauthorized and disruptive parties.

“Airbnb is committed to supporting hosts and the communities where they live, and we hope these defenses will allow guests, hosts and neighbors to celebrate the holidays with more confidence.”

The company said the measures will be implemented in countries and regions around the world in the lead-up to and throughout the new year, and restrictions will apply to bookings of one to three nights.

In addition, guests making reservations are required to confirm that they understand that unauthorized and disruptive parties are not permitted on Airbnb and, if found to be policy is violated, will be suspended from the platform.