Yelp claims "no human programmed these results"
Yelp searches are perpetuating the racist stereotype that Asian American restaurants serve cat and dog meat.
According to Tampa Bay Times, Yelp searches are suggesting that Asian American restaurants are serving cat and dog meat.
Searching for dog meat, Yelp suggested Korean restaurants whilst a search for cat meat would result in Chinese restaurants being suggested.
Such results were found in searches in a dozen U.S. cities and regions, including Tampa Bay, Chicagoland, the San Francisco Bay area, New York, Philadelphia, Denver, Boston, Austin, Seattle, Atlanta and Jacksonville.
No reviews for Korean restaurants contained the words “dog meat”. The only reviews that featured the word “dog” in Korean restaurant reviews were in the context of it being pet-friendly or referring to giving their dog leftovers.
Similarly, restaurant reviews for Chinese restaurants rarely contained the word “cat”.
A spokesperson for Yelp said searches rely on, “real-world consumer user data and human behavior patterns. That means keywords found in Yelp’s user-submitted restaurant reviews, plus users’ previous searches and behaviors on the app can affect them.
“Included in the huge volume of search queries Yelp receives are some very rare, atypical ones that computer-generated models still try to match … which is made more difficult by the rarity of this search occurrence,” the statement added. “To be clear, no human programmed these results or matches and we are taking prompt action to remove them from autocomplete and our other systems.
“Thank you for bringing this to our attention and allowing us to correct it.”
Ken Lee, CEO of OCA (Organization of Chinese Americans), called Yelp’s search engine an example of “algorithmic bias”.
“Our small businesses must already combat racist, inflammatory reviews from users on Yelp. These biased search results for dog and cat meat vendors should not be an additional concern,” Lee said. “The company has a social and corporate responsibility to their consumers and local businesses to keep their app and website clear of prejudice and misinformation.”
“Algorithmic bias against communities of color continues to plague new technology and online platforms — this incident is another in a long line of incidents.”