By Harry Kemsley OBE — This case looks like it’s about social media harm. It is more revealing than that. It shows how influence can be built into the environment itself, shaping what feels normal without ever needing to persuade you directly.
Why read this essay?
- Understand how influence has shifted from content to system design — and why that matters.
- Learn how platforms shape behaviour through patterns, not individual posts.
- Gain practical ways to recognise and respond to influence in your own feed.
What is this case really about?
At first glance, the case against Meta and YouTube appears to be about a familiar question: does social media harm young people?
But the legal arguments point to something more specific.
The plaintiff did not simply argue that social media correlated with poor mental health. The claim was that these platforms were designed in ways that make certain behaviours more likely, especially in vulnerable users.
Features such as infinite scroll, autoplay, notifications, filters, and visible metrics like ‘likes’ and ‘followers’ were presented not as neutral tools, but as systems that shape behaviour over time.
The defence responded with a different framing. They argued that mental-health outcomes are multi-causal, “addiction” is not clearly defined in clinical terms, and individual life circumstances matter more than platform design.
Both positions contain truth.
But the more important question for Luminae is not who “won” the argument. Instead: what does this case reveal about how influence actually works now?
From persuasion to ‘shaping the environment’
Traditional ideas of influence focus on persuasion. Someone presents an argument. You agree or disagree. But the mechanisms described in this case operate differently. They do not rely on convincing you.
They rely on shaping the conditions in which your choices are made.
- Infinite scroll removes natural stopping points.
- Autoplay reduces the need to choose.
- Notifications create anticipation and return behaviour.
Over time, this shifts behaviour without requiring conscious decisions at each step.
So, the key shift is that influence is no longer primarily about what you are told. It is about what you are repeatedly exposed to, and how that exposure is structured.
Why patterns matter more than posts
One of the easiest mistakes to make is to evaluate influence at the level of individual content.
- “Is this post misleading?”
- “Is this video harmful?”
But influence rarely operates through a single item. It emerges through patterns. A feed that repeatedly shows aspirational lifestyles, productivity systems, specific body ideals, or particular worldviews does not need to tell you what to think. It gradually shifts what feels normal, what feels desirable, or what feels possible. This is why the case focused on design features rather than specific pieces of content.
The argument is not that one post causes harm. It is that a system of repeated exposure can shape direction over time.
When influence aligns with your identity
One of the more subtle insights from the case is that influence does not need to work against you to be effective. It often works through you. Algorithms learn from behaviour: what you watch, what you linger on, or what you return to.
From this, they build a profile — not of who you are in a full sense, but of what you are likely to engage with. Then they reinforce it. Why? Because that is how they measure success: maximising engagement.
For example, if you are drawn to mastery, improvement, identity-building, or aspiration, the system does not need to push outrage or fear. It can instead amplify those same dimensions. This makes influence harder to detect, because it feels aligned.
The role of vulnerability and timing
A critical piece of evidence in the case was internal research suggesting that some users are more susceptible than others — particularly those experiencing stress, transition, or uncertainty.
This matters because influence is not evenly distributed. It is often strongest during periods of change, when identity is still forming, or when emotional states are more variable. At those moments, repeated patterns of content can normalise certain paths or behaviours, make some choices feel “obvious”, and crowd out alternatives.
This does not require intent in a conspiratorial sense. It can simply emerge from engagement optimisation. But the effect is the same: the system adapts faster than the user becomes aware of it.
How influence hides in plain sight
The most effective influence does not feel like influence. It feels like “this is what lots of people are doing”, or “this is what makes sense now”, or “this is just where things are going”. Luminae refers to this as unexamined influence: not forced, nor explicit, but shaping direction nonetheless.
Look at your recent or current information feeds as you scroll through and ask yourself a useful question: if your feed had an agenda, what would it be? Not in a dramatic sense, but in terms of what behaviours it encourages, what identities or emotions it reinforces, and what choices or opinions it makes more likely.
What does ‘agency’ look like in this environment?
The goal is not to reject platforms or avoid influence entirely. That is neither realistic nor necessary. The goal is to see the system clearly enough to make conscious choices within it.
A few practical ways to do that:
- Notice patterns, not just posts. Ask: why am I seeing so much of this category of content?
- Track desire drift. Compare what you want now with what you wanted a month ago. What changed?
- Reconstruct the path backwards. If something feels like a new “obvious” idea, ask where it came from.
- Introduce small friction. Pause before acting. Search intentionally instead of scrolling. Turn off non-essential notifications.
- Use the third-person test. If someone else saw your feed, what would they think it was trying to get you to do?
These are not defensive tactics. They are awareness tools: slowing the information environment, reducing or at least recognising the noise, and regaining agency.
So what?
You do not need to decide whether social media is good or bad. But you do need to recognise that your environment is not neutral. What you see repeatedly shapes what feels normal. What feels normal shapes what you do next.
The shift is simple, but important:
From — “This is what I’m interested in”
To — “This is a pattern — do I want to follow it?”
Shine a light on the unexamined influence — that is where agency lives.
References
AP News (2026), Closing arguments begin in landmark social media trial set in Los Angeles; CBS News (2026), Social media companies accused of “addicting the brains of children”; CBS News (2026), Plaintiff testimony in social media addiction trial; The Guardian (2026), Are infinite scroll and autoplay creating addicts?; Bloomberg Law (2026), YouTube lawyer argues usage patterns do not support addiction claim.