 |
| Twitter is testing a new appeal process for sensitive content |
Twitter has launched a new test of an updated appeals process for violating sensitive content to allow users who believe they are being unfairly punished through the platform's visual detection algorithm to find such decisions.
This process provides an easier way to handle requests for content from Tweets. If the user content is flagged as an error, it helps the user to get a response from the platform.
The image recognition process of this platform has become the cause of many notable misclassifications.
The platform struggled last month after tagging a photo of a burnt veteran in military uniform on Twitter on July 4 for potentially sensitive content.
Twitter should rely on AI learning as much as possible to limit the spread of harmful content. His system leads to improvements in this area.
In a recent transparency report, the platform reported that during that period, the number of impressions of offending tweets (tweets that broke the rules for some reason) was less than 0.1% of the total number of tweets.
Twitter is testing a new appeal process
Over time, AI-based identifiers can help reduce the visibility of tweets significantly.
The platform also reported a 322% increase in the number of sensitive content policy violations and a 194% increase in the number of accounts responding to nudity.
Although these results are generally good. But this increased activity and use of preventative automatic restraint systems can also lead to more false positives.
For this reason, this improved complaints procedure makes perfect sense. It is hoped that the platform can improve the detection system while meeting the needs of users who have been misjudged.
Initial tests of the new recovery process will be made available to a small group of users in Mexico, the United States and Japan.
The platform said: If you are participating in an experiment, please do not mark your content as sensitive in the settings. You can object to Tweets being reported from the Tweet itself (entire network) or alert notifications.
This update is in line with the broader evolution of the platform towards improved user protection.