The state of social media in 2025

Twitter was rebranded as X in July 2023 and has become part of Musk’s plan for an “everything app”. One of the key changes that Musk made was to pivot from advertising to paid subscriptions for generating revenue. This resulted in the previous verification system – which was a reliable method for confirming the identity of accounts belonging to businesses, prominent figures and journalists – being replaced by one that promotes content for whoever pays (with some exceptions) for the premium service on X.

X also downsized its workforce and now relies on a community notes feature for fact-checking. While crowdsourcing information can work, such as with Wikipedia, community notes can be slow and misinformation is not always reliably identified. Nonetheless, similar practices have recently been adopted by Meta for its Facebook and Instagram platforms, as it has stopped using independent fact-checkers.

With the removal of fact-checking, social media bots have become a widespread problem. These are often intended as a propaganda tool and can artificially create or promote content that algorithms will showcase. Algorithms, which curate the content users see, will often share posts from people who users do not follow or have any interest in. For example, even if a user does not follow Elon Musk on X, his posts will still be frequently promoted to them.

“I think the attempt to monetise the platform has had what I call ‘rage farmers’ increase enormously; you can monetise saying awful, evil things, because they’re going to get so many more retweets and comments,” says Hanson. “My advice would be to not set up a Twitter [X] account, unless you are concerned that somebody may hijack your identity.”

As a consequence of recent policy changes, the promotion of far-right content and the training of AI, many users are choosing to leave X

Many social media companies are now training AI using content published on their platforms by users. Meta has Meta AI, while X has Grok. Although some platforms, such as Facebook, allow users to opt out of their information being used in this manner, X does not give its users this option. The training of AI is part of some social media companies’ wider plans to shift into the AI market.

The training of AI is especially concerning for those working in the creative fields, as any designs, artwork or writing could be incorporated into the training of the platform’s AI. This has become a key driver for artists and writers deciding to move away from using X.

New social media platforms have also emerged. Meta has launched a microblogging platform called Threads. Meanwhile, decentralised social networks, such as Mastodon and BlueSky – the latter initially developed as a research project at Twitter before becoming independent – have become popular.

The anonymising nature of social media has emboldened some users to post aggressive, extreme or harmful content. Many social media users feel there has been an increase in far-right rhetoric. “It’s worse than ever; I don’t think I’ve ever seen so much deep vitriol and aggressive violent language,” says Hanson.

Following Donald Trump’s inauguration as US president, Meta updated its guidelines to say: “We do allow allegations of mental illness or abnormality when based on gender or sexual orientation, given political and religious discourse about transgenderism and homosexuality and common non-serious usage of words like ‘weird’.” This change in policy has left some users feeling less safe than they did previously.

Social media migration

As a consequence of recent policy changes, the promotion of far-right content and the training of AI, many users are choosing to leave X. One of the most prominent departures was The Guardian, which stopped using X, citing “disturbing content promoted or found on the platform”.

Tania Stevenson, an independent councillor in the Derby area, is currently petitioning local authorities and public bodies to disengage from using X. The petition states that X has become “increasingly associated with hate speech, intolerance, and bigotry”.

Many users are migrating to newer platforms, such as Threads, Mastodon or BlueSky. However, there is much more to migrating to a new social media site than creating a new profile. Users need to let their followers know they are leaving the platform. Deleting or deactivating an account is not recommended, as usernames can be made available for other users after 30 days, which could lead to social media cloning.

Some have gone as far as to delete all their posts from a platform to prevent their content from being used to train an AI. Although platforms do not typically provide tools for bulk deletion, scripts can be used, such as this one for X/Twitter, which can mass-delete posts.

Therefore, to prevent cloning, users should consider maintaining their existing social media profiles, but indicating that they are no longer using them by directing followers to other platforms where they can be found.

Verification and identity on social media

Confirming identity continues to be a challenge on social media, with profiles being cloned to impersonate other people for a variety of reasons, including fraud. This issue has been compounded by the decision to allow new profiles to be created using the usernames of deleted accounts.

As a consequence of X changing the verification system to a premium service, it no longer has the valued recognition it once did. Meta offers blue tick verification, but it is marketed as a subscription service to promote content rather than as a tool for confirming identity.

BlueSky has taken a different approach to verification. Instead of confirming the identity of a user, BlueSky offers a system whereby a profile can confirm ownership of a website. Most BlueSky users will have “bsky.social” after their username, but by inserting a few lines of code into the user’s website, a BlueSky profile name can match the domain name of a website. This solution is by no means flawless, but it does make cloning those accounts more difficult.

The future of social media

The social media landscape continues to undergo a paradigm shift, as organisations and individual users question whether they wish to continue to support certain platforms.

Twitter/X has been consistently losing users since Musk bought it and started changing policies. Facebook, which is predominantly used by older generations, is also showing declining growth, but Instagram remains popular.

Ultimately, if something is free, then the user is the product. In exchange for using any social platform, every user generates income for the platform, through advertising and subscriptions that can be used for whatever the owner decides

Threads is showing active use, but not to the degree of other recent platforms. This may be partly due to the algorithm being programmed to downgrade political content, to avoid outrage. Mastodon is gaining popularity, but there have been some challenges due to the distributed nature of the platform. However, BlueSky continues to grow and has quickly proven popular for certain communities, especially in the creative fields. As of late February 2025, there were over 31 million users of BlueSky. The online publishing platform Medium is also popular for hosting long-form posts, such as blogs.

“I would definitely set up a BlueSky account, because the forum value on BlueSky is enormous and, in terms of targeted marketing, you can be hyper-specific with it,” says Hanson. “I’d set up a Medium account, because that way you have total control over what your blog posts are and it’s much harder to say something stupid in 600 well-thought-out words, and you’ll have an established place for your opinions and audience growth.”

Ultimately, if something is free, then the user is the product. In exchange for using any social platform, every user generates income for the platform, through advertising and subscriptions that can be used for whatever the owner decides. For some platforms, this can include user data being used to train their AI.

Many mainstream social media platforms are currently facing a trust issue. Users are questioning whether they want their posts to be used to train an AI. Meanwhile, far-right commentators are emboldened to post inflammatory content for engagement value, while there is an increase in various types of disinformation campaigns due to insufficient content moderation.

As such, users are questioning whether they wish to continue being associated with certain platforms if they believe they have become associated with hate speech or engage in exploitative behaviour when training their AI. As such, users are engaging with alternative social media platforms and ensuring they no longer rely on just a single social media outlet.