Tumblr is the latest technology company to cope with the automated labeling and the collapse that has passed and put the anger of users.
In recent days, Tumblr users have complained that their content is highlighted as “mature”, even when this is not the case. The problem has reduced the viewing of users’ posts because many people on the platform have set their settings to hide mature content by default.
According to numerous publications by users influenced by Tumblr, publications have falsely highlighted despite the fact that they are not sexual or violent and these false collapse have included everything from Cat GIF in Fandom content to art and even one image of hands. Some suspect that AI -based automation could blame the issues.
Similar categories have contaminated other social media platforms in recent weeks. For example, Pinterest finally admitted that an internal error led to the bans of the mass user. Meanwhile, Instagram this week has refused to comment on its own problem with the massive bans stated by users have been paid little coverage or attention beyond electronic complaints in applications such as X and Reddit.
In both cases, users suspected that moderation based on AI was responsible, although Pinterest denied that it happened.
In Tumblr, the label issue is linked to an update in the Android application, where the company stated that it is experimenting with improvements in mature filtering systems. Specifically, it tests a new layer of moderation Content labelsThe company told TechCrunch.
A Tumblr spokesman said the experiments are continuing and, according to user comments, will improve before the changes are circulated on other platforms.
“As we work to make Tumblr a safer place for all users, we aim to respect a different set of content interests and preferences that can be adapted to the settings. We see this as a continuing process as we continue to refine the way we find it and treat it.
The company too recognized the problem in his blogResponding to a user’s question about falsely marked positions. Here, Tumblr’s team noted that they knew “wrong classification issues” and were actively working to reduce them.
In addition, the post explained that the Tumblr appeal process was informed in the coming weeks to be able to handle a higher volume of cases. (Tumblr did not answer our questions about scheduled changes in the appeal process.)
Whether the AI or the other automation is blame is not clear, as Tumblr will not speak specifically about the cause of the new issues. However, reduced staffing in the blogging service probably also played a role.
Following 2019 by WordPress.com’s Maker Automattic, Tumblr faced layoffs as its staff was transferred to other works of the parent company. Last year, Automattic announced that the back of the Tumblr will also move to WordPress to facilitate management and stop its financial losses.
