A fresh report from Mozilla, the makers of the privacy-focused Firefox browser, shows that YouTubes user controls are ineffective at controlling what folks see on the platformdespite what Google claims. Using data from almost 23,000 volunteers, Mozilla could show that YouTube kept recommending similar videos even though people used the many different options to point they didnt desire to see that sort of content.
YouTube may be the second-most popular website on earth (the foremost is Google) and in accordance with Mozilla, around 70 percent of the 1 billion hours viewed daily on the platform are because of algorithmic recommendations. Various reports show the way the algorithm can polarize people and recommend misinformation and harmful contenta thing that Google claims it spent some time working hard to repair. In this study, Mozilla attempt to test the potency of the controls YouTube offers to users to control the recommended videos they see.
In a previous report released in July this past year, Mozilla discovered that individuals were routinely recommended videos they didnt desire to see and felt that the controls open to them were ineffective. This new study used a browser plug-in Mozilla developed called RegretsReport to see if this is true.
Mozilla viewed four different Google-suggested controls: Clicking that thumbs-down Dislike button, Not interested, Dont recommend channel, and Remove from watch history. Meanwhile, users of the RegretsReport plug-in visit a Stop Recommending button on YouTube videos. If they clicked it, the control-option (such as for example for the Dislike button) corresponding with their test group was delivered to YouTube, while data about future recommended videos were delivered to Mozilla. (There is also a control group where clicking the button did nothing.)
[Related: Why YouTube is hiding dislikes on videos]
During the period of the analysis, 22,722 participants used the RegretsReporter, allowing Mozilla to investigate 567,880,195 recommended videos. To assess this large amount of data, the researchers reviewed 40,000 pairs of recommended videos and rated their similarity. This allowed the team to quantitatively study if the videos participants were being recommended were much like videos they had previously rejected. Put simply, to check out whether YouTubes tools effectively reduced the amount of bad recommendations.
For instance, if someone saw an anti-vax video recommended in their mind, and clicked Not interested, and got recommended a cat video, that might be an excellent recommendation. However, should they kept getting suggested anti-vax videos after indicating they werent thinking about them, those will be bad recommendations. Page 22 of the report [PDF] has the right visual examples.
Mozillas report discovered that no user control was especially able to preventing unwanted recommendations. The Dont recommend channel option had the largest impact, preventing 43 percent of bad recommendations, with Remove from watch history preventing 29 percent, and Dislike rather than interested preventing 12 percent and 11 percent, respectively. Mozilla argues that its research shows that YouTube is not actually that thinking about hearing what its users want, preferring to depend on opaque methods that drive engagement whatever the needs of its users.
Due to its findings, Mozilla is contacting visitors to sign a petition asking YouTube to repair its feedback tools and present users actual control on the videos they get recommended. In addition, it has four specific tips for YouTube and policy makers predicated on its study.
Mozilla shows that YouTubes user controls ought to be simple to use and understand, and become made to put people in the drivers seat. In addition, it wants YouTube to grant researchers better usage of data (so that they dont need to use browser extensions to review these types of things). Finally, it calls on policy makers to pass laws providing legal protections for all those engaged in public areas interest.
Whether this report will do to obtain Google to include some real user controls to YouTube remains to be observed. For the present time, its a reasonably damning indictment of the ineffective controls which are currently set up.