Twitter removed 300K terror-related accounts in first half of 2017, a 20% decline from previous six months, with 75% suspended before posting first tweet
BRUSSELS (Reuters) - Twitter Inc (TWTR.N) said that its internal controls were allowing it to weed out accounts being used for the …
Context & Ripple Effects
Twitter's disclosed enforcement totals had risen from 125,000 terrorism-related suspensions since mid-2015 to 235,000 more accounts removed in the following six months, before the company reported more than 636,000 removals through the end of 2016. The new disclosure adds a quality measure to those volume reports: much of the enforcement occurred before an account could publish.
The subsequent report of 274,000 terrorism-promoting accounts suspended in late 2017 indicates that Twitter continued reporting removals at a large scale even as it emphasized earlier intervention.
First-order effects
- Terror-related accounts identified by Twitter in the first half of 2017 lost access to the platform, with three quarters stopped before a first post could reach users.
- Twitter can point to pre-publication suspensions as evidence that its internal controls are acting before harmful accounts become visible on the service.
Second-order effects
- For Twitter, the 20% decline in removals is less informative as a standalone measure when a larger share of enforcement happens before posting; prevention rate becomes a more consequential operating metric alongside total suspensions.
Third-order effects
- If Twitter sustains pre-post detection, terrorism enforcement shifts from removing visible content and accounts toward screening account creation, making the platform's internal detection systems central to moderation outcomes.
The trend: Platform safety enforcement is moving upstream from reactive takedowns toward preventing suspected abusive accounts from becoming active.