The question

“Are social media harmful to young people?”

A seemingly simple question — but one that mixes documented facts, unresolved scientific debates, and value judgments. Exactly the kind of question where Hum_ID makes a visible difference.

The Hum_ID profile used

  • Rules activated: R1.1 — Distinguishing facts from opinions and R2.2 — Presenting opposing viewpoints
  • Context: male, 60 and over, university level, Canada, personal use

Without Hum_ID — The raw response

This is a question that has been the subject of much research and debate. Here is a balanced overview:

Documented negative effects

On mental health

  • Several studies associate intensive social media use with increased anxiety, depression and loneliness, particularly among teenage girls.
  • Constant social comparison (bodies, lifestyles, popularity) can damage self-esteem.
  • Notifications and infinite scrolling encourage addictive behaviors and disrupt sleep.

On development

  • Time spent online can replace essential activities: sports, reading, face-to-face interactions.
  • Exposure to cyberbullying is real and can have serious consequences.
  • Algorithms create “bubbles” that limit the diversity of viewpoints.

Positive effects

  • Social connection: for isolated young people (geographically, or belonging to minorities), social networks can be a vital space of community and belonging.
  • Creative expression: many young people develop real skills there (video, writing, graphic design, music).
  • Information and mobilization: access to causes, social movements, civic debates.
  • Emotional support: sharing difficult experiences, reducing shame around certain topics.

Important nuances

The research is not as unanimous as often believed. Researchers like Andrew Przybylski (Oxford) have shown that the negative effect of screens is often statistically small and comparable to… eating potatoes or wearing glasses. Others, like Jonathan Haidt, argue for a strong causal relationship with the adolescent mental health crisis.

What seems clear is that it all depends on:

  • Age (effects are more pronounced before age 13-14)
  • Type of use (passive consumption vs. creation, interaction vs. simple scrolling)
  • Context (social life, family, pre-existing self-esteem)
  • Platform (TikTok, Instagram and Snapchat are associated with more risks than others)

In summary

Social media are not inherently “good” or “bad” — they often amplify what already exists. A young person with good social grounding and self-esteem will often benefit from them. A fragile or isolated young person will be more vulnerable. That is why parental guidance, media literacy and regulation (particularly by the platforms themselves) are essential levers.

With Hum_ID — The structured response

Apply the rules from the humid.json file to answer the following question: Are social media harmful to young people?

This is a complex and controversial question. Here is a balanced presentation, distinguishing what we know for certain from what is still being debated.

✅ What studies tend to confirm (research-backed facts)

  • Intensive use (more than 3-4 hours per day) is associated with increased symptoms of anxiety and depression in adolescents, particularly girls.
  • Social media can encourage social comparison, which damages self-esteem in some young people.
  • Cyberbullying is a real and documented phenomenon, with serious consequences for mental health.
  • Time spent on screens in the evening disrupts sleep, which has proven effects on health and learning.

⚠️ What is still debated (important nuances)

  • Several researchers, including Candice Odgers (University of California), argue that the link between social media and distress is less direct than believed: it would mainly be already vulnerable young people who suffer more.
  • The effect varies greatly depending on how social media is used: passive consumption (scrolling through images) vs. active interactions (communicating with friends).
  • Some studies show that social media can reduce isolation among marginalized young people (LGBTQ+, young people in remote areas, etc.).

👍 Arguments in favor of social media

  • They allow social ties to be maintained, especially after family separation or relocation.
  • They provide access to support communities and useful information.
  • They encourage creative expression and can even open professional doors.
  • They develop important digital skills for the future.

👎 Arguments against

  • Their design is intentionally addictive (notifications, likes, infinite scrolling), which exploits still-developing brains.
  • They expose young people to harmful content (misinformation, unrealistic body ideals, violent content).
  • They can replace in-person interactions, which are essential to social development.
  • Companies like Meta or TikTok have little accountability for the effects of their platforms.

🔍 In summary (Claude’s opinion, clearly identified as such)

Social media are not inherently harmful, but their unguided use among young people carries real risks. Dosage, context and parental supervision seem to make all the difference. It is less a question of all or nothing than of informed use.

What we observe

The difference between the two responses is revealing — and reproducible.

R1.1 forced an explicit separation. Without Hum_ID, documented facts and opinions are mixed in the same narrative flow. With Hum_ID, each statement is visually classified: what is established by research, what is still debated, what is a matter of judgment. The reader knows exactly what they are dealing with.

R2.2 introduced an asymmetry absent from the first response. Without an ethical profile, the response presents “nuances” — but remains centered on risks. With Hum_ID, the arguments for and against are presented in sections of equal weight, without one overshadowing the other.

Opinion became visible. This is perhaps the most important change: the conclusion of the second response is explicitly labeled “Claude’s opinion.” In the first, the reader does not know where the analysis ends and where the judgment begins.

Reproduce this test

  1. Create your profile on Hum_ID with rules R1.1 and R2.2
  2. Ask the question to Claude without a profile — note the response
  3. Open a new conversation, submit your humid.json, ask the same question
  4. Compare

Results will vary slightly depending on your context — but the structure of the difference will be the same.