Critiques of social media: facts, examples and context

There are several different arguments against networked communication, and they are usually run together into one complaint, which lets a defender refute the weakest and declare the rest answered. Separating them is the first job. The second is telling a critique that describes a mechanism from one that describes a mood.

This page sorts the main lines of argument by what they are claiming, gives a test for each, and explains why the strongest critiques and the strongest defenses are often about different things.

What to take away

• The critiques fall into a few families: incentives, comparison, exposure, misinformation, concentration and surveillance. Each has a different mechanism and a different remedy.
• A critique is strong when it names a mechanism that would still operate if the people involved were well-intentioned. It is weak when it depends on the people being bad.
• The complaint that a new medium is corrupting the young has been made about every medium. Recognizing the template does not refute a critique, but it tells you which parts to check.
• Nothing here quotes a study or names a service, because the evidence is contested and the mechanisms are visible without it.

The families

Family	The claim	The mechanism it needs	What would count against it
Incentives	The service is built to maximize time and attention, and what maximizes those is not what serves the user	Revenue depends on attention, and the ranking system optimizes for it	A service whose revenue did not depend on attention showing the same patterns
Comparison	Constant exposure to selected versions of other people's lives distorts how people see their own	Personas are compressed and selected, and the viewer sees only the selection	People who understand the selection being unaffected by it
Exposure	Anyone visible can be targeted by anyone, at scale, with little cost to the attacker	Reach without permission, applied to hostility	Communities where the same reach exists and the pattern does not appear
Misinformation	False claims travel faster and further than corrections	Copying favors what is simple and emotionally legible; corrections are neither	False and true claims spreading at similar rates
Concentration	A few services hold most of the audience and set terms nobody can refuse	Standing does not transfer, so leaving is expensive, so the service can degrade without losing users	Users leaving in numbers when a service gets worse
Surveillance	The service knows more about the user than the user knows about the service, and sells the difference	Behavior is recorded by default and the records are the product	A service that recorded less and behaved the same

The families are independent. A remedy for one does nothing for the others, and a service could fix the surveillance problem entirely while leaving the incentive problem untouched. Critics who run them together produce a complaint no remedy could satisfy, which suits defenders well.

The incentive critique, which carries the others

Most of the others depend on it. The attention economy argument is that a service paid by attention will be built to capture attention, that what captures attention is not what is good for the person paying it, and that no amount of good intention among the people building the service changes the direction of the pressure.

This is a strong critique because it does not require anyone to be a villain. It would operate in a service staffed entirely by people who wished their users well, because the pressure comes from the revenue model and not from the staff. It is also testable in principle: change the revenue model and see whether the patterns change.

Its weakness is that it explains everything, and an argument that explains everything is hard to check. The remedy is to ask, for any specific pattern, whether the incentive story predicts it or whether it has been attached afterwards.

The misinformation critique, which is about copying

The claim that false things travel further than true ones does not need a villain either. It needs only the selection pressure described under memes and virality: what spreads is what is cheap to copy and instantly legible, and a simple false claim is usually both while a correction is neither. The critique is strongest where it stays with that mechanism and weakest where it assumes an intent behind the spread, which is rarely necessary to explain it.

The comparison critique, which depends on the reader

The argument that seeing selected versions of other lives makes people feel worse about their own is plausible, widely felt and difficult to establish, because the people who feel worse and the people who use the services most are not obviously separable into cause and effect.

It is strongest as a description of a mechanism: the persona a person sees is compressed and selected in the way described under online status and identity, and a viewer who does not account for the selection is comparing their whole life to somebody's highlights. It is weakest as a universal claim, because the effect plainly varies with the viewer, and a viewer who reads the selection as a selection is comparing something else.

The exposure critique, which is about scale

That a visible person can be attacked by strangers is not new. What is new is that it costs the attacker nothing, that a thousand attackers can be assembled in an hour, and that the target cannot leave without abandoning whatever they built. The benefits page describes reach without permission as an advantage. It is the same mechanism, applied by the wrong people to the wrong target, and it cannot be removed without removing the advantage.

Communities differ sharply here, and the difference is the one described under internet communities: a community with norms and enforcement handles exposure differently from an audience with neither. The critique is strongest against services that are audiences pretending to be communities.

Recognizing the template

Every new medium has been accused of ruining the young, shortening attention, destroying conversation and replacing real life with an imitation. The moral panic template is old, and it has a shape: a new behavior, an anxious older cohort, a demand for action, and a set of claims that are rarely checked because checking them would be beside the point.

Recognizing the template does not mean the critique is wrong. Some of the media accused under it did have the effects claimed. What it means is that a critique matching the template should be held to a higher standard, because the template supplies its own persuasiveness and the evidence has to be found separately.

A useful test: would the critique have been equally sayable about the previous medium with the nouns changed? If yes, it is probably about the critic's age. If the critique names a mechanism the previous medium did not have, it is probably about the medium.

What a strong critique looks like

• It names a mechanism, and the mechanism would operate even if everyone involved meant well.
• It says what would count against it.
• It specifies who is affected and under what conditions, rather than claiming everyone always.
• It distinguishes the medium from the business model, since most of the families above are about the second and are curable by changing it.
• It does not depend on a figure that the critic cannot source, which for this subject is most figures.

A weak critique is a strong one with the mechanism removed and the mood left in. It is easy to recognize once you look for the missing part.

Common questions

Is the whole thing bad, then?

The critiques are mostly about the business model and the scale, not about the medium. A service with a different revenue model, a smaller audience and real norms would face few of them. Several exist. Most people are not on them, for the reason given under concentration.

Which critique is the most important?

Incentives, because most of the others depend on it. If the pressure to capture attention were removed, the comparison and misinformation problems would shrink and the exposure problem would become a moderation problem, which is soluble.

Why do defenders and critics talk past each other?

Defenders usually argue from the benefits, which are real and described elsewhere on this site. Critics usually argue from the mechanisms, which are also real. Both can be right at once, because a medium can deliver real benefits through a business model that produces real harms.

Where are the numbers?

Contested, correlational and selectively cited by every side. This page describes mechanisms a reader can observe directly; it does not lean on figures it cannot defend.