Netflix Inc. has responded to viewers accusations that its onsite recommendations are tailor made to match the race of the viewer.
Several Netflix viewers complained via Twitter that Netflix populated thumbnails of titles featuring African American actors, despite them not playing the lead roles.
The Netflix Originals movie Like Father was touted as an example by several Twitter users. The film stars two caucasian actors, Kelsey Grammer and Kristen Bell, but ads were populating showing the film’s minor characters played by African American actors Leonard Ouzts and Blaire Brook.
I rarely watch "Black" titles on Netflix in comparison to the rest of the content I actually view. But I DO feel to unnecessarily targeted to watch what their algorithm thinks I should be watching. That may be why I purposely don't watch Black flicks on NF.
— Michael-Alan (@MikeInBmore) October 19, 2018
Other titles such as Love Actually, The Good Cop and Set It Up were also used to demonstrate the somewhat unusual ads.
Netflix firmly denies these accusations with the following statement: “Reports that we look at demographics when personalizing artwork are untrue.
“We don’t ask members for their race, gender or ethnicity, so we cannot use this information to personalize their individual Netflix experience. The only information we use is a member’s viewing history.”
Just to add my white data point, these are the thumbnails I see for some of the same movies you shared. pic.twitter.com/GjapwcdQuD
— Kelly Quantrill (@codetrill) October 19, 2018
Netflix unrolled its new algorithm to match tailor made posters of recommended content to its users. In a Netflix Inc company blogpost, it demonstrated how different movie posters would be matched to a customer’s viewing history.
It uses movie posters for Goodwill Hunting as an example, showing different posters that would appear for a romantic comedy fan versus a viewer with a history of watching regular comedy films.
Tim Harrington, a senior broadcast analyst told The Guardian that Netflix’ new algorithm method was ‘downright clunky.’
“Netflix’s recommendation engine is second to none, and works almost seamlessly in the background.
“But the algorithm for targeted artwork is shown as downright clunky when, say, [some] users are offered artwork for ITV’s Lewis with black actors despite both leads and almost the entire cast being white.
Netflix knows a lot about you, perhaps even race, but their understanding of what to do with this information is currently rudimentary,” Harrington said.