@[email protected]

To our partners, we want to address a recent Bloomberg article which claims that harmful content viewed on Twitter has gone up over the past six months. This is an absolutely false assertion. 99.99% of Tweet impressions are healthy. Which means only a tiny amount of content requires enforcement. But we’ll keep doing whatever we can to make this platform as safe and healthy as possible. Bloomberg created this article using outdated research that contains incorrect or misleading metrics. The article does not properly provide the right context or new updates to the remediations we have made since the third party research was conducted. Here’s more information: Incorrect Metrics: - The third party researchers consistently discuss "tweets authored’ rather than the "volume of the tweets" that are actually seen. That is a very important distinction. - Twitter's move to Freedom of Speech, Not Reach (blog.twitter.com/en_us/topics/p…) has dramatically reduced the reach of ‘lawful and awful’ Tweets by an average of 81% compared to healthy Tweets. - Twitter's work is not done, but we’re encouraged by the progress in driving hate speech impressions 30% lower on average compared to pre-acquisition. Outdated and Old Data: - Much of the data within the report is from the period of time immediately after the acquisition. - Which have already been reviewed or actioned accordingly through automated and manual enforcement. - Since acquisition, Twitter has shipped products that have led to significant improvements not limited to, the aforementioned work, and brand safety and suitability tools that validate () that our efforts are more than 99% effective in placing ads adjacent to safe content Zero Tolerance: - Twitter has zero tolerance (help.twitter.com/en/rules-and-p…) for content that incites violence, threatens people violence, or wishes harm on people. - Under our Abuse & Harassment policy ( ), Twitter restricts sharing of abusive content, engaging in the targeted harassment of someone, or inciting other people to do so. - This includes behavior that is repeated, unreciprocated, and intended to humiliate or degrade an individual(s) with abuse or hate. Furthermore, Twitter has strong brand safety and suitability controls in place that are effective in keeping your ads safe. Our brand safety engineering team is also actively building more suitability capabilities focused on inventory filtering and pre-bidding that will be available in two weeks. Current Brand Safety Capabilities: - Brands can choose which surfaces (timelines, replies, search results, immersive video) their content will appear on - Twitter offers first party Adjacency Controls that enable customers to create a negative keyword and account handle list that will prevent their ads from appearing adjacent one spot above or below tweet content they have signaled is unsuitable in Twitter’s Home Timeline, where the majority of all consumption occurs - Twitter currently partners with industry leaders Integral Ad Science (IAS) and DoubleVerify (DV) to power “post-buy” measurement for Brand Safety & Suitability. Brand Suitability Capabilities: - Twitter is actively building inventory filtering capabilities that will allow even more control over content adjacencies. - Twitter is exploring new pre-bidding control options with third parties. Topline Points - 99.99% of Tweet impressions are healthy. - 5X increase in account suspensions for Child Sexual Exploitation vs. a year ago - 95% of suspensions are proactive–before we get a user report — up from 75% a year ago Advertising Related Data - As observed by our third party partners, 99%+ of measured ad impressions appeared adjacent to content that was deemed safe in accordance with the GARM brand safety floor. - For advertisers leveraging Adjacency Controls, we’ve observed a 99% efficacy rate in avoiding terms and authors brands deem inappropriate. - More than 1,800 customers are now using adjacency controls. In summary, we’re committed to providing you with a safe and suitable environment where you can connect with your customers. We’ll continue to enhance our suite of brand safety solutions to allow for even greater transparency and control. And we’ll continue to demonstrate through data that our platform is safe and is getting safer every day.