From Twitter’s misinformation bird hide, we risk not seeing the forest for the trees


Many months of being at home, and the popularity of birdwatching has “soared”. Soon, however, we can also engage in this beloved activity online and tune into the not-always-so-sweet melodies to be found on Twitter.

Initiatives like the Big Garden Birdwatch are a method of collective citizen data-gathering: relying on local communities, because they have access to information others don’t. Twitter’s new ‘Birdwatch’ operates on similar principles. A “community-driven approach to address misinformation” being piloted first in the US, with Birdwatch users having to take the following steps:

  1. See a tweet they believe is misleading 
  2. Add a note explaining why and what they believe the correct information is instead 
  3. Have their note rated by other users on how helpful it is, and ranked in line with this

There can be a tendency to approach misinformation moderation in the spirit that content is either true or false, and which one it is can be easily decided and acted upon by a platform. In short, this doesn’t work. As our research has repeatedly highlighted, there are significant complexities of how information can be harmful or benign, while being true, false, both or neither. Birdwatch, encouragingly, moves away from simple tick-box moderation.

And Twitter is honest about some of the challenges: potential for manipulation and harassment, and bias based on the distribution of contributors’ “ideology, background, or interest space”. There’s cautious optimism: Evan Greer, the director of online activists Fight for the Future, has said Twitter “should consult with experts and members of impacted communities” as they “experiment with possible solutions” but “it makes sense to explore more decentralized models”.

Despite this, it may struggle to leave its pilot nest. The question of how users are compensated for their Birdwatching hasn’t been answered. The incentives are the promise of making Twitter a better space and the attraction of ratings, but with a “community” of over 330 million inhabitants, it’s questionable whether a representative group will feel interested or invested (or safe) in volunteering.

More fundamentally, it’s right that consultation and decentralisation are explored. But Birdwatch is not decentralised; Dorsey still rules the roost. It is the project of a commercial platform where users are not stakeholders with genuine representation and major influence. Birdwatch is open to consultation but, in addition to still being beholden to Twitter’s rules, there is no accountability for users’ feedback going ignored or reneged upon. 

All this speaks to the need for even greater democratisation. As we laid out in a recent paper, content moderation failures arise when:

“[…] unaccountable, authoritarian regimes set and enforce profit-maximising rules on powerless populations that lack the means or incentives to build out the pro-social structures and cultures upon which a healthy, democratic society depends.”

Recently, Twitter also announced ‘Bluesky’: a developing “decentralized standard for social media” which it will “ultimately be a client of”. But clouds gathered when critics argued these already exist and so this was simply a power grab. Be it watching the trees or building in the skies, it all feels a little like Twitter is willing to open up but only on its own terms.

Similar to Birdwatch is Wikipedia, which struggles with editor diversity – and experts have highlighted why responding to this requires continued democratisation. Yet, being a non-profit, Wikipedia is perhaps fundamentally more able to address concerns over misinformation and its relationship to its user base in this way.

Birdwatch is a worthwhile experiment, but it’s still about tackling misinformation on Twitter’s terms. Ultimately, we might need to look beyond its hide to spot the solution.