Twitter’s testing out a new element in its push to limit the spread of false and misleading content in the app, with a variable labeling system that would alert users about the latest updates and info around topical subjects within tweets, while also providing links to authoritative resources on the same.

As you can see in this example, posted by reverse engineering extraordinaire Jane Manchun Wong, the new labeling system would include three different types of in-stream alerts on flagged tweets, in many cases triggered by specific keywords included.

Those labels, at this stage of development, are:

• ‘Get the Latest’ – This would apply to fast-evolving news topics, providing Twitter with a means to connect users to authoritative info on any tweets including specific words, like, for example, ‘election’ or ‘poll’, which would then give users an immediate connection to the latest reports from official sources.
• ‘Stay informed‘ – This could also be applied to evolving news topics, though maybe for more sustained news stories, like COVID-19, again with links to connect people through to key authorities.
• ‘Misleading’ – The last category would be ‘misleading’, which would clearly flag tweets including questionable claims, and provide links, again, to official sources.

The first two, as noted, could likely be triggered by keywords included within tweets, but the last would probably need to be checked by Twitter’s moderation teams before being flagged, as it could be difficult for Twitter’s automated system to detect the relevant context.

It could be an effective way to provide more information around evolving topics – but then again, it could also clutter user timelines, depending on how much the labels are applied, and what, exactly, triggers their inclusion.

In this respect, it would likely come down to Twitter’s discretion, with the platform only applying the alerts to limited, chosen topics. Which could make it a viable, helpful tool – but it’s still early days, and it’s difficult to predict the full potential impact without all the relevant details on hand.

Twitter has confirmed that this new system is currently in testing, and that it’s iterating the best application for the new labeling options. Some users have noted that these alerts look very similar to quote tweets, which could be confusing, while others have suggested that more prominent coloring for each kind of alert could be a better way to highlight these tweets in-stream.

Twitter’s still looking for the best way forward, but it is interesting to note the latest developments, and how Twitter’s working to evolve its misinformation detection and alert process.

And before you note – yes, this would be based on Twitter’s existing fact-checking processes, which means that it is indeed Twitter’s own moderation teams and third-party connections that decide what would qualify under each label. If you don’t believe that Twitter can be trusted to make the right call on such, I’d suggest using another platform, but Twitter does work with a range of third-party authorities and groups to ensure it connects users through to the right information on each topic.