Twitter Says It Now Removes Half Of All Abusive Tweets Proactively

It attributes the progress to its machine learning capabilities continuing to improve. But prominent politicians are still permitted to spread falsehoods at will. JL

Paul Sawers reports in Venture Beat:

Twitter said it now removes half of all abusive tweets proactively, without relying on anyone to report them.  Relying on user reports and human moderators to remove abuse is a near-impossible challenge on a platform of Twitter’s scale, which is why it has turned to machine learning tools to automate much of the process. This increase is largely attributable to “improving our machine-learning models in Q3” to detect potential policy violations, even though these policies don’t always apply to everyone equally.

Twitter said it now removes half of all abusive tweets proactively, without relying on anyone to report them.
Twitter has faced an uphill struggle as it strives to make its platform a less daunting place to spend time — bullying and abuse are genuine threats for anyone daring to engage with others on Twitter. But relying on user reports and human moderators to remove abuse is a near-impossible challenge on a platform of Twitter’s scale, which is why it has turned to machine learning tools to automate much of the process, just as Facebook is doing.

As part of today’s Q3 earnings report, which saw Twitter grow its “monetizable daily active users” (mDAUs) by 6 million compared to the previous quarter, the company said 50% of all tweets it removes are handled automatically, without users having to report the abuse.
“We continue to make progress on health, improving our ability to proactively identify and remove abusive content, with more than 50% of the tweets removed for abusive content in Q3 taken down without a bystander or first-person report,” Twitter cofounder and CEO Jack Dorsey said.
For comparison, Twitter reported this figure at 38% for Q1 this year and 43% for Q2.
According to the company, this increase is largely attributable to “improving our machine-learning models in Q3” to detect potential policy violations, even though these policies don’t always apply to everyone equally.
“We will continue our work to proactively reduce abuse on Twitter, with the goal of reducing the burden on victims of abuse and, increasingly, taking action before abuse is reported,” the company wrote in its letter to shareholders.
Twitter has made a number of changes to its platform in recent times, as it looks to encourage businesses and the public to hang around. Last month, the company rolled out a controversial “hide replies” feature that allows users to curate responses to their tweets, raising the risk that high-profile figures such as politicians will whitewash responses to dubious claims they make.
Earlier today, Twitter reported Q3 2019 revenue of $824 million, missing estimates by around $50 million, with its mDAUs jumping to 145 million. The company’s shares fell 20% in premarket trading.