According to Reuters, the new head of trust and safety at Elon Musk’s Twitter stated that the company is heavily relying on automation to moderate content, preferring distribution restrictions over outright language removal.
Ella Irwin, Vice President of Trust and Safety Product at Twitter, stated that despite the potential effects on “benign uses” of those terms.
In the first interview a Twitter executive has given since Musk’s acquisition of the social media company at the end of October, Irwin stated on Thursday, “The team is totally empowered to move quickly and to be as aggressive as possible, which is the largest shift..”
After Musk announced an amnesty for accounts suspended under the company’s previous leadership that had not violated the law or engaged in “egregious spam,” researchers have reported a surge in hate speech on the social media service.
Since Musk slashed half of Twitter’s workforce and issued an ultimatum to work long hours, the company has faced pointed questions about its ability and willingness to moderate illegal and harmful content.
Additionally, concerns regarding brand safety have driven away advertisers, Twitter’s primary source of revenue.
During a meeting with French President Emmanuel Macron on Friday, Musk promised “significant reinforcement of content moderation and protection of freedom of speech.”
Irwin said that Musk told the team that safety was the company’s top priority and that they should not worry about how their actions would affect user growth or revenue.
According to former employees who are familiar with the work, the safety approach that Irwin described at least partially reflects an acceleration of the changes that were planned for Twitter’s handling of hateful conduct and other policy violations since last year.
One strategy, which is summed up in the industry adage “freedom of speech, not freedom of reach,” entails excluding tweets that go against the company’s policies from the home timeline and search.
Before the Musk acquisition, Twitter had already included such “visibility filtering” tools for misinformation into its official hateful conduct policy.The method reduces the risk of harm caused by abusive content that goes viral while increasing speech freedom.
According to the Center for Countering Digital Hate, the number of hateful tweets on Twitter increased significantly in the week before Musk tweeted on November 23 that impressions, or views, of hateful speech were decreasing. This is one example of researchers pointing to the prevalence of such content while Musk claims a reduction in visibility.
According to the findings, there was a 31% increase in tweets containing a gay slur and a triple increase in tweets containing words that were anti-Black that week.
‘MORE Dangers, MOVE Quick’
Irwin, who joined the organization in June and recently held wellbeing jobs at different organizations including Amazon.com and Google, pushed back on ideas that Twitter didn’t have the assets or eagerness to safeguard the stage.
She stated that full-time employees and contractors working in “critical areas” like child safety and content moderation in the company’s “Health” division were not significantly affected by the layoffs.
According to two sources with knowledge of the cuts, more than half of the Health engineering unit was laid off.Although he had previously denied that layoffs had a significant impact on the Health team, Irwin did not immediately respond to a request for comment regarding the assertion.
She added that the team’s product manager was still employed and that the number of people working on child safety had not increased since the acquisition.Despite her refusal to provide specific figures regarding the extent of the turnover, Irwin stated that Twitter backfilled some positions for employees who left the company.
She argued that the company had erred in the past by using time- and labor-intensive human reviews of harmful content, and that Musk was focusing more on using automation.
Regarding child safety, for instance, Irwin stated that Twitter had moved toward automatically removing tweets that were reported by reputable individuals who had a track record of accurately flagging posts that were harmful.
Carolina Christofoletti, a threat intelligence researcher at TRM Labs who focuses on content that depicts child sexual abuse, claims that in recent times, Twitter has removed some content without either acknowledging receipt of her report or providing confirmation of its decision.
In the interview on Thursday, Irwin stated that Twitter and the cybersecurity organization Ghost Data collaborated to remove approximately 44,000 accounts associated with child safety violations.
The search results for “teen” pornography and other hashtags that are frequently associated with abuse are also being restricted on Twitter.She stated that previous concerns regarding the impact of such restrictions on permitted uses of the terms had subsided.
According to Irwin, the utilization of “trusted reporters” had previously been “something we’ve discussed at Twitter, but there was some hesitancy and frankly, just some delay.”