According to Chloe Colliver, head of the digital research unit at the ISD think tank, the perceived political biases must be analysed on two fronts: the algorithms at the heart of social media firms and the people shaping the decisions.
There is not enough data, she says, to suggest the codes that determine which posts get flagged as egregious are inherently biased, though she recommends creating a sandbox in which these alogrithms can be tested.
But things get more complicated when it comes to decision makers. “If you look at individual decisions being made on political leaders, the response by companies has overwhelmingly been sporadic and patchy across the political spectrum,” Colliver says.
It is a perceived issue that Trump has highlighted frequently, with his supporters often singling out individuals for alleged bias.
The latest target is Yoel Roth, Twitter’s head of site integrity. In recent days, Roth has received death threats after social media users uncovered posts he sent criticising the president.
“From their bogus ‘fact check’ of @realDonaldTrump to their ‘head of site integrity’ displaying his clear hatred towards Republicans, Twitter’s blatant bias has gone too far,” tweeted Republican National Chairman Chair Ronna McDaniel.
This isn’t a new accusation. Last year the White House unveiled a “Tech Bias Reporting tool” meant primarily to highlight cases in which conservatives believed they were being silenced
A year before that, non-profit firm Freedom Watch filed a lawsuit that accused the likes of Twitter, Google and Facebook of favouring liberal voices.
The two-year legal battle ended in failure this week after the claims were dismissed. Despite the loss, many other allegations of liberal bias remain.
US senator Josh Hawley called a conservative-led review of alleged bias at Facebook a “smokescreen” after the report concluded that it had found no evidence of bias inside the company.