Twitter is launching a new feature on iOS that encourages users to review “potentially harmful” messages
Stop and think before sending that tweet! Twitter is relaunching a feature on iOS that encourages users to review “potentially harmful” responses
- Twitter has relaunched a feature that encourages users to review tweets
- If the AI detects malicious language, it will prompt you to review the context
- The rollout is currently an experiment on iOS and can be ignored by users
Twitter asks users to think before publishing “a potentially harmful answer.”
The social media platform announced on Tuesday that if offensive words are detected in an answer by the AI, the system will ask users to consider and review the context.
The feature is currently an experiment on iOS and those involved in the test will receive a notification allowing them to ‘revise’, ‘delete’ or ‘send’ the tweet.
Twitter ran similar variations to the feature last year when the AI was deemed capable of detecting malicious language in tweets – and it looks like the social media giant is trying again.
Twitter asks users to think before publishing “a potentially harmful answer.” The social media platform announced on Tuesday that if offensive words are detected in an answer by the AI, the system will ask users to consider and review the context
In May 2020, the feature was rolled out to iOS to convince users to think before releasing a potentially harmful tweet.
“When it heats up, you might say things you don’t mean,” reads May’s announcement.
“To get you to rethink an answer, we’re running a limited experiment on iOS with a prompt that gives you the option to revise your answer before publishing if it uses language that could be harmful.”
However, the feature quietly disappeared and resurfaced a few months later, in August 2020, on Android, iOS and the web, but was canceled again from the platform.
In May 2020, the feature was rolled out to iOS in an attempt to convince users to think before releasing a potentially harmful tweet
And it seems like the third time is the charm as Twitter is giving the feature one more shot.
“Say something when you might regret it,” Twitter shared in Tuesday’s announcement.
We have relaunched this experiment on iOS asking you to review an answer that may be harmful or offensive. Do you think you accidentally received a prompt? Share your feedback with us so we can improve. ‘
According to the social platform, it is conducting a test for users on iOS who will use its AI to scan replies before posting and give users a chance to review or revise them.
Users are also free to ignore the warning message and post the response anyway.
The problem Twitter can face, like others who use AI moderators, is that the system isn’t always correct and can be biased when flagging tweets.
A Cornell University study found that tweets believed to be written by African Americas are more likely to be tagged as hate speech than tweets associated with whites.
Researchers found that the algorithms classified African American tweets as sexism, hate speech, harassment, or abuse at much higher rates than the tweets believed to have been written by whites – more than twice as often in some cases.
The researchers believe the inequality has two causes: an oversampling of African-Americans’ tweets when databases are created; and insufficient training for the people who annotate tweets for potentially hateful content.
And it seems like the third time is the charm as Twitter is giving the feature one more shot
Twitter has rolled out a host of features over the past year aimed at fighting hateful speed and misinformation on the platform
YouTube recently had another problem with its AI moderator when it shut down a channel focused on chess.
Last summer, a YouTuber who produces popular chess videos saw his channel blocked from including what the site called “ harmful and dangerous ” content.
YouTube didn’t explain why it had blocked Croatian chess player Antonio Radic, otherwise known as ‘Agadmator’, but service was restored 24 hours later.
Carnegie Mellon computer scientists suspect that Radic’s “ black versus white ” discussion with a grandmaster accidentally triggered YouTube’s AI filters.