What counts as 'social media'? Aston study finds the public can't say
Aston University researchers say the term 'social media' has no fixed meaning in public discourse — and that the platforms least agreed-upon are the ones doing the most identity work.
On 10 July 2026, researchers at Aston University published findings that cut against a decade of confident assumptions about the internet's most-studied object. The term "social media," they reported, has no settled public definition, and the platforms that people disagree most about are precisely the ones that sit at the centre of contemporary platform governance fights.
The implication is awkward for regulators in Brussels, Westminster and Washington who have spent years drafting rules premised on a category that, the data now suggests, ordinary users do not actually share. If "social media" is a moving target, so is the boundary of every law drafted in its name.
The finding, plain
The study, reported by Phys.org, surveyed members of the British public and found no broadly accepted working definition of "social media" — neither a technical one, nor a feature-based one, nor a usage-based one. Respondents disagreed on which sites and apps belong in the category. The disagreement was not random. Familiar services like Facebook, Instagram, TikTok and X were grouped together by most respondents. Less familiar or recently rebranded platforms produced the most variance. People could not agree on whether YouTube, WhatsApp, Reddit or LinkedIn qualify.
That matters because the platforms sitting in the no-man's-land of contested classification — messaging apps, video hosts, professional networks, discussion forums — are also the ones most frequently the subject of safety legislation, age-verification rules and content-moderation mandates. The category ambiguity, in other words, lives exactly where the policy fights are loudest.
Why the gap exists
Three structural pressures keep the definition fractured. First, the platforms themselves have refused a stable taxonomy: parent companies rebrand, fold products into each other, and ship features across categories. WhatsApp introduced channels; Instagram added text threads; TikTok pushed longer video; LinkedIn launched a feed. The product surface keeps migrating across the very lines a clean definition would require.
Second, usage has migrated with it. A platform can be "social" in the marketing sense while functioning in practice as search infrastructure (YouTube), broadcast media (a single viral creator), customer service (a brand's DMs), or a marketplace (Facebook Marketplace). The technical object and the lived use have drifted apart.
Third, the political economy of the term rewards imprecision. Platforms prefer to be called "social" when the designation buys them Section 230-style liability shelter in the United States or lighter duties under the EU's Digital Services Act, and prefer to be called something else — "messaging," "streaming," "professional" — when the designation would attract heavier obligations. The same company will use different labels in different jurisdictions depending on which rulebook it wants to be filed under.
The regulator's problem
European Union rules on illegal content, transparency and systemic risk are anchored to a category — "very large online platform" or "very large online search engine" — that presupposes users can tell which is which. The United Kingdom's Online Safety Act uses "user-to-user" and "search" services, again presupposing a clean boundary. Australia's age-assurance regime targets "social media platforms" by name. India's IT Rules use "social media intermediary" and "intermediary" as distinct categories with different obligations.
If the public cannot reliably place a given service inside those categories, compliance officers at the companies cannot reliably argue either. The contested cases will be the ones that matter: the encrypted messenger with public channels, the video host with a social graph, the professional network that has become a news feed.
What the ambiguity costs
The most concrete cost is to the people the rules are written for. When a regulator cannot say clearly which platforms its rules cover, users cannot know which protections apply where they actually spend their time. A teenager whose primary internet is TikTok, Discord and YouTube lives inside a patchwork where the rules governing each piece were drafted assuming the others did not exist.
The second cost is to the platforms that play by the spirit of the rules. Companies that have invested in moderation, transparency reporting and risk assessments have an interest in the category being narrow enough to exclude freeloaders — services that behave identically but claim a different classification to escape the same duties.
The third cost is to the research base. A decade of academic, journalistic and policy work on "the effects of social media" rests on a category the public cannot reproduce. That does not make the work wrong, but it does mean the headline numbers — time spent, mental-health correlations, exposure-to-misinformation estimates — sit on a denominator that has never been pinned down.
A counterweight
The simplest objection is that definitional messiness is normal. "Television" was contested for decades. "News" still is. The law operates on functional categories, not on folk categories, and reasonable regulators can be expected to write tests that capture a service's function rather than its label. The Online Safety Act's "user-to-user" test is itself an attempt at a functional definition.
The rejoinder is that functional tests still have to be administered in public, and the public cannot meaningfully consent to a regime whose category boundaries it cannot recognise. Democratic legitimacy, not legal craft, is the resource being spent.
What to watch next
The test case will be the next contested designation in the EU's DSA enforcement, where the Commission has to decide whether a specific large service is a "platform" or something else. The UK media regulator Ofcom is expected to publish its first set of categorisation decisions under the Online Safety Act later this year. Each of those calls will be, in effect, a public answer to the question Aston has now documented: ordinary people asked, point blank, did not know.
The Aston study is a small survey of a single country. Its contribution is not to settle the definitional fight but to demonstrate, with data, that the fight is real and that the public sits on the losing end of it. The platforms least agreed-upon are the platforms least regulated, and the platforms least regulated are the ones the public cannot even name consistently.