Twitter’s Birdwatch: ‘The Misinformation Effect’ in Reverse
Twitter has announced a new Wikipedia-style fact-checking feature, called Birdwatch.
The idea is to develop a community of people who can help annotate Tweets that contain questionable material. Twitter’s research finds that adding this context is more helpful to users than applying a ‘True’ or ‘False’ label.
Birdwatch allows people to identify information in Tweets they believe is misleading and write notes that provide informative context. We believe this approach has the potential to respond quickly when misleading information spreads, adding context that people trust and find valuable.
Twitter will make all Birdwatch data available for public download. There is a pilot version live here, in the US only.
There will be a little icon below Tweets, which users can click to report the content to the Birdwatchers for investigation.
Jane Manchun Wong saw an early version of the report form in October:
This requires the user to consider why they are reporting the content. There are thoughtful distinctions here, such as “Satire/joke that may be misrepresented” and “Outdated information that may be misleading”.
Next, the complaint will be assessed by the Birdwatching community, with their conclusion and any actions taken available for public view.
It is a well-intentioned scheme with a couple of significant considerations:
- Is Twitter simply outsourcing responsiblity for handling misinformation on the platform?
- Does it go far enough?
Twitter is asking for users to help in policing content and this may be an honest assessment of their own position today. After all, should a company like Twitter decide what is misleading for the rest of us?
Yet the company must take swift action against the more egregious abuses of its technology. One imagines there will be a division of responsibilities here between Twitter and its Birdwatchers.
The second question is most pressing in the immediate term, although it impacts the former quite directly.
Twitter applied labels to some of Donald Trump’s Tweets, before removing the former president from the platform. For example, it showed that Trump’s claims about electoral fraud were “disputed”.
Twitter also put restrictions in place so that users could not spread Trump’s disinformation any further.
Misinformation is false information shared regardless of intention to deceive; disinformation is shared with the express intention of deceit.
I bring these cold definitions to bear so we might discuss the ‘misinformation effect’, before analysing how it might help Twitter tackle disinformation.
The misinformation effect occurs when post-event information shapes our memory of the actual event. A study found that after people witness a car crash, their memory is altered by subtle deviations in the questions they are asked after seeing the crash.
Participants in the study watched the same video. Then they were split into groups and asked one of the following two questions:
- “How fast were the cars going when they hit each other?”
- “How fast were the cars going when they smashed into each other?”
Then they were asked, “Did you see broken glass?”
Those who were initially asked the “smashed into” version of the question were much more likely to report having seen broken glass.
In this scenario, we begin with the objective facts as displayed in the video clip. The facts then become ‘truth’ as they are perceived and absorbed by the individual observer. Truth, for all its lofty connotations, is today the stuff of subjectivity.
Misinformation enters the fray with the deft substitution of “smashed into” for “hit”.
And it has an impact. The new information is easier for us to recall and it frames our recollection of the initial facts.
Might the same logic apply in reverse?
What if we start with the disinformation shared on Twitter, then counter this with a factual label?
Might that modify the truth the reader creates, bringing it closer to the facts?
There remains a lingering suspicion that the damage is already done. Our default mode is to believe what we read; it requires too heavy a mental load to scrutinise everything we see. A label may dampen the impact of disinformation, but it will not nullify it altogether.
We must also consider that disinformation is only one means to an end for Twitter’s bad actors.
As we have discussed here before, Donald Trump used Twitter to great effect — if not always with great cunning.
He lied. But as the famous formulation goes, some people took him literally while others took him seriously.
His opponents balked at his claims about “building the wall”.
His supporters weren’t too concerned with the minutiae of the wall’s construction; they just saw a president who was finally ready to take action on immigration.
Calling out the factual errors is playing chess while the real recipients of the message play checkers. The facts are not the real game here.
In the post-Trump world we can expect a new breed of populists, emboldened by Trump’s rise to power. They do not share a deep-seated love of lying; lying is just a good way to get the job done.
If Twitter’s labels do decelerate the spread of disinformation, they will find other means.
For example, how would you label this Tweet from Ted Cruz?
Well it is factually inaccurate, sure. But one suspects that Cruz is very well aware of this already.
So why post it?
The real reason may lie in the reaction it receives.
His opponents’ gut reaction is to seize the opportunity to mock Cruz.
Because, how could you resist?
Cruz then doubles down by sharing a picture of the bizarre bumper stickers he’s selling on his website:
Cruz wants “the libs” to mock him, to show their elitism. He knows his followers do not see him as an unintelligent man and the sub-text of his statement will resonate with them. When they see the other side round on him, it strengthens their existing prejudices.
No doubt, Twitter views Birdwatch as just one part of a broader strategy to tackle disinformation. Any action on this front is to be welcomed and it is an ongoing battle.
However, there are strategists who spend all their waking hours working on schemes to use social media to their employers’ ends. If Twitter starts applying labels to posts soon after they are published, politicians will move into a hieroglyphic world of symbols best understood by their fans.
These grifters should be taken seriously, even when they are acting obtuse.
Taking them literally will only play into their hands.