ARTICLE AD BOX
By Shayan Sardarizadeh & Rachel Schraer
BBC Monitoring & BBC Reality Check
What makes one social media post pop up all over people's feeds and another disappear into obscurity? It's a question anyone trying to make a splash online would love the answer to.
We know that tweets don't just appear in chronological order - your feed is being engineered in some way. And platforms tend to keep the full details of their algorithm - the code that determines which posts you see in what order - under their hats.
Under an initiative by Twitter's new chief Elon Musk, fragments of internal Twitter discussions about content moderation - dubbed "the Twitter Files" - have been shared on the platform by a few selected journalists.
The ongoing revelations contain internal chats of Twitter's inner circle of senior executives about a series of controversial decisions: limiting the dissemination of the Hunter Biden laptop story, the permanent suspension of Donald Trump in the wake of the 6 January 2021 Capitol riot, and limiting the reach of some influential accounts.
Part two of the revelations, shared by US journalist Bari Weiss, who had been provided with screenshots by Twitter's new head of safety Ella Irwin, reveals three influential conservative accounts in the US were put on different "blacklists", meaning their tweets either wouldn't trend, might not appear in search, or would not be "amplified".
Ms Weiss said decisions were made "all in secret, without informing users". Mr Musk went even further, saying the revelations proved "the rules were enforced against the right, but not against the left".
But it's incredibly hard to verify whether this is giving us the whole picture.
We are missing the context of which other accounts have faced similar treatment, and whether the restricted accounts were in breach of other rules, for example inciting hate or spreading false claims about Covid that could cause harm.
Restricting accounts can be a useful tool if they are spreading harmful material. A common phrase when we talk about regulating the internet is "freedom of speech not freedom of reach" - it's one Mr Musk himself has used to describe the platform's new approach.
Twitter, for instance, confirmed just a few days ago that it would not "amplify tweets containing slurs or hate speech".
The idea of an account being placed on "trends blacklists" doesn't seem totally out of step with this policy. The more controversial part comes from who gets to decide what can and can't trend, where the threshold for a restriction lies and who ends up on the list.
One name on the Twitter files list that has drawn attention is Dr Jay Bhattacharya - a health policy professor at Stanford University - who vocally opposed coronavirus lockdowns, and whose tweets were reportedly prevented from trending.
He has been criticised, including by a judge in a US case about mask mandates in schools, who accused him of oversimplifying research studies, and by other scientists for proposing an alternative to lockdowns they said was unworkable and would lead to more deaths.
But he has said science required people to disagree with each other, arguing against "censorship of scientific discussion".
There have been various reports suggesting marginalised groups including trans and plus size people were more likely to have their accounts restricted.
Multiple liberal and left-leaning individuals have also previously complained about their accounts and tweets being "suppressed".
The difficulty is, people will have a hunch they are being restricted- the engagement with their posts suddenly falls or their followers report not seeing them anymore - but that could be down to many other reasons, and it's incredibly hard to prove that's what is happening without access to the companies' internal workings.
Alex Stamos, director of the Stanford Internet Observatory and a former Facebook chief security officer, says Mr Musk is "providing extremely limited transparency" to like-minded US media figures. He proposed that Mr Musk should make available details of all communications around content moderation with governments, political parties and politicians around the world.
Mr Musk says he intends to introduce a new feature that will enable all users to see "your true account status", including whether they've been "shadowbanned" (hidden from other people's feeds), the reason for it, and a process to appeal.
Twitter isn't the only company moving in this direction- Instagram recently launched a new tool of its own to let you know if your posts are barred from being recommended to other users.
While these changes add some degree of transparency to the process, content moderation decisions are incredibly difficult and messy, and remain open to political influence from owners, senior executives and pressure from politicians and activists.
The BBC approached Ms Weiss and Twitter for comment.
Analysis
By Marianna Spring, Disinformation and social media correspondent
At its heart, this is a row over whether social media companies like Twitter should control what we're exposed to on their sites - and how transparent they are about how they do that.
The 'Twitter Files' are challenging how the company made decisions about enforcing certain policies - and dealing with certain accounts.
But, former employees I've spoken to are quick to point out that much of what they tell us is not new and that ultimately, how you interpret these 'Twitter Files' depends on where you stand on how social media sites should deal with misinformation and hate.
We already know about Twitter's policies to tackle both of these issues, some of which involved suspending accounts as well as removing content.
Twitter's curation teams, axed since Musk bought Twitter, were in charge of deciding which content could trend and what would be de-prioritised, including on issues like misinformation and discriminatory hate.
After the pandemic, the riots at the Capitol in the US and the war in Ukraine, Twitter came under huge pressure to do more about these issues.
What's new about the Twitter Files is detail on who was making the decisions - and specific accounts that have been targeted.
Revealing internal conversations has sparked hate and backlash online directed at those caught up in this. And there are still unanswered questions about how other accounts have been affected, not just the select few listed.