{"id":31284,"date":"2023-07-13T16:42:44","date_gmt":"2023-07-13T16:42:44","guid":{"rendered":"https:\/\/danielsmarketingblc.com\/CuratedProducts\/?p=31284"},"modified":"2023-07-13T16:42:45","modified_gmt":"2023-07-13T16:42:45","slug":"twitter-claims-incredibly-high-performance-on-content-moderation-in-contrast-to-third-party-reports","status":"publish","type":"post","link":"https:\/\/danielsmarketingblc.com\/CuratedProducts\/2023\/07\/13\/twitter-claims-incredibly-high-performance-on-content-moderation-in-contrast-to-third-party-reports\/","title":{"rendered":"Twitter Claims Incredibly High Performance on Content Moderation, in Contrast to Third-Party Reports"},"content":{"rendered":"

<\/p>\n

\n

Okay, we\u2019re going to get asked about this, which is the one reason I\u2019m going to cover it \u2013 because really, it\u2019s not possible to look at such claims with none access to the info in query.<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/p>\n

Today, Twitter tweeted out a brand new claim<\/a> that \u2018<\/span><\/span><\/span><\/span><\/span>greater than 99.99% of Tweet impressions are from healthy content, or content that doesn’t violate our rules\u2019. <\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/p>\n

As per Twitter<\/a>:<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/p>\n

\u201cOn Twitter individuals are free to be their true selves. On a regular basis, we work to preserve free speech on Twitter, while equally maintaining the health of our platform. For the reason that launch of <\/span><\/span><\/span><\/span>\u2018<\/em>Freedom of Speech Not Reach\u2019, we\u2019ve seen encouraging results. That is why in the approaching weeks, we\u2019re expanding the applying of this enforcement motion from our <\/span><\/span><\/span><\/span>Hateful Conduct policy\u00a0to now include our policies on <\/span>Abusive Behavior\u00a0and <\/span>Violent Speech.\u201d<\/span><\/em><\/span><\/span><\/span><\/p>\n

Twitter\u2019s \u2018<\/span><\/span><\/span><\/span>Freedom of Speech Not Reach\u2019 approach, which it outlined back in April<\/a>, essentially explains that Twitter\u2019s now looking to cut back the reach of some less violative content within the app, despite it technically breaking its rules, versus removing such outright. Twitter has also added labels to those tweets, to make clear when such motion has been initiated.<\/span><\/span><\/span><\/span><\/span><\/span><\/p>\n

<\/div>\n

As Twitter notes, originally, this approach was only applied to tweets that will previously have been deemed in violation of its <\/span><\/span><\/span><\/span>Hateful Conduct policy, however it\u2019s now trying to expand this same systematic enforcement motion to abusive and violent tweets as well.<\/span><\/span><\/span><\/span><\/span><\/span><\/p>\n

To be clear, Twitter\u2019s rules around such content haven\u2019t modified, but its enforcement approach is different, in that previous Twitter management would have removed more of these kind of comments outright – but now, Twitter\u2019s taking a more lenient approach, by reducing their reach as a substitute.<\/span><\/span><\/span><\/span><\/span><\/span><\/p>\n

And Twitter says that that is working \u2013 with a staggeringly low 0.01% of tweets that violate its rules now being seen by any users in any respect.<\/span><\/span><\/span><\/span><\/span><\/span><\/p>\n

Which seems not possible, based on overall industry trends, and external reporting on Twitter specifically.<\/span><\/span><\/span><\/span><\/span><\/span><\/p>\n

For instance, the prevalence of comparable violations on Facebook on Instagram sits at around 0.05%<\/a> – and Meta has much more staff, and much more advanced systems working to handle such across its apps. The suggestion that Twitter has someway been capable of best this, after culling 80% of its staff<\/a>, including lots of the individuals who were working on addressing these elements, seems questionable at best.<\/span><\/span><\/span><\/span><\/span><\/span><\/p>\n

There have also, as noted, been a variety of third-party evaluation reports which suggest, for instance, that <\/span><\/span><\/span>antisemitic tweets have develop into more common<\/a> since Elon Musk took over on the app, that slurs against Black and transgender people have also increased<\/a>, while hate speech, generally, has also develop into more prevalent<\/a> amid the app\u2019s broader changes in approach. Twitter\u2019s also facing legal motion in each Australia<\/a> and Germany<\/a> for failing to remove hate speech in a timely manner<\/span><\/span><\/span><\/span><\/span>.<\/span><\/span><\/span><\/p>\n

As we\u2019ve reported previously, a few of the conflicting figures here seem to return right down to various definitions of what actually qualifies as hate speech<\/a>, and the way Twitter itself is measuring such. But we don\u2019t understand how Twitter has come to this recent 99.99% figure, because there\u2019s no evidence – the Twitter team hasn\u2019t provided any actual data or insight to back this number up. <\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/p>\n

So it\u2019s just \u2018take out word for it\u2019, that someway, Twitter has achieved record-setting results sparsely performance, despite cutting the vast majority of its staff, and in contrast to external academic evaluation, which points to the other.<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/p>\n

I\u2019m not saying that it\u2019s not right, but I don\u2019t know, and also you don\u2019t know either, because Twitter hasn\u2019t explained itself in any way. <\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/p>\n

So what meaning, ultimately, I don\u2019t know. <\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/p>\n

But sure, it\u2019s a powerful figure, I suppose.<\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/span><\/p>\n<\/div>\n