Twitter is pushing a change to timelines, to have an 'algorithmic' timeline of 'relevant' content rather than a straightforward record of content as submitted by users. Twitter is also, due to pressure from the US Gov't, claiming to use anti-spam techniques to quash speech by 'terrorist' groups. These two items are probably linked.
Twitter is said to be introducing an algorithmic timeline and is marketing such as a means to make the Twitter experience more relevant to people. This particular change has garnered significant backlash amongst many in the Twittersphere, as they dislike the notion that the views that they've chosen - the view based on a specific list of folks that they follow - would be processed to change the content away from the user's specific choice of who to follow and replaced with Twitter's estimation of what is more 'relevant'. The notion that tweets might be filtered or replaced with more 'relevant' (read: generating more income, such as the 'promoted' tweets) is anethema to many, especially those who suspect that such usage may be intended to quash inconvenient or controversial points of view.
If you're just displaying a timeline of the posts people make, any changes are obvious censorship. Change that and "It's just the algorithm" - @munin
It's very important to note that algorithms are written by people, and written for a specific purpose. If something is 'filtered' by an algorithm, this means that someone either purposely ensured that it would be removed or deprecated, or that they had no desire to ensure that point of view was represented. Algorithms are a means of reflecting an organization's purpose - they are not some magical way to absolve oneself from responsibility.
Computers do what they are programmed to do - they do not act on their own.
Keeping this in mind, we can also take a look at the other significant Twitter story happening concurrently: Twitter is blocking 'ISIS' accounts presumably due to pressure by the US Government. This concurrency is slightly suggestive - while not absolute proof, it is a potentially interesting coincidence that 'algorithmic' timelines (that can be used to quash unwanted points of view) and pressure to quash certain politically unwanted speech are happening at the same time.
Keep in mind also that Twitter is a company - it is beholden to the interests of its shareholders, who would be presumably unhappy if the company was unable to make money due to litigation accusing it of aiding 'terrorist' organizations.
Put these things together and it isn't terribly surprising that Twitter is rolling out these changes - changes which allow them to sculpt the direction that discourse takes on their platform, and allows them to satisfy the demands that have been made by people with the ability to curtail their operations.
Twitter was never really a 'free' platform for its users. The users are, much like every other "Web 2.0" property, the commodity that the company was selling - generally to advertisers. The real owners of Twitter - its investors - want it to continue as a going concern; the way for them to do so with the least amount of friction is to accede to the demands that are being made that they censor certain speech. To censor speech obviously would cause an outcry; therefore, it would be plausible that they would hide such censorship in a new 'feature' to make the discourse more 'relevant' - and here we are.
Needless to say, I am in no way connected to Twitter other than having an account there, and this is merely my opinion as a highly suspicious and cynical user. Take this with a grain of salt - and never trust what you read, especially on social media.