this post was submitted on 12 Dec 2025
15 points (94.1% liked)

Canada

10752 readers
693 users here now

What's going on Canada?



Related Communities


🍁 Meta


🗺️ Provinces / Territories


🏙️ Cities / Local Communities

Sorted alphabetically by city name.


🏒 SportsHockey

Football (NFL): incomplete

Football (CFL): incomplete

Baseball

Basketball

Soccer


💻 Schools / Universities

Sorted by province, then by total full-time enrolment.


💵 Finance, Shopping, Sales


🗣️ Politics


🍁 Social / Culture


Rules

  1. Keep the original title when submitting an article. You can put your own commentary in the body of the post or in the comment section.

Reminder that the rules for lemmy.ca also apply here. See the sidebar on the homepage: lemmy.ca


founded 5 years ago
MODERATORS
 

As of Wednesday, all youth under 16 in Australia will be banned from major social media platforms like TikTok, Instagram, Snapchat, YouTube, Reddit, Twitch, and X. For over a decade, whistleblowers, politicians, academics, and experts around the world have sounded the alarm about the online harms people of all ages are exposed to.

...

The ban does nothing to prepare teens to respond to digital harms. It makes no investments in education, community training, or parental support. Youth will not be magically prepared to address problematic online behaviours or content when they turn 16.

The time and resources spent on the ban could be better spent on things like providing education and support for digital citizenship, media literacy, privacy rights or resource centres.

If social media is problematic for a 13, 14 or 15 year old, it’s still likely to be problematic for a 16, 25, or 80 year old. There is no body of research that establishes 16 as a “safe threshold” for social media use and the age for healthy use can vary across genders.

...

Under the current model, companies will not be inclined to improve their reporting systems for harmful content. In fact, in response to the ban, YouTube is actually removing a feature that would allow teens to report content they find inappropriate.

Youth under 16 who find ways to use these platforms, despite the bans, will be unlikely to come forward and ask for help if things go wrong. After all, they weren’t supposed to be online in the first place.

The answer to mitigating online harms is not kicking teens offline.

...

Social media companies also need to be accountable to the ways the platforms are designed and run. These platforms are designed in ways that push certain content and elicit particular engagements.

...

top 14 comments
sorted by: hot top controversial new old
[–] jaselle@lemmy.ca 14 points 9 hours ago (2 children)

None of these are good arguments against introducing a ban. Worst argument of all is that "we shouldn't ban it for 15 year olds because that wouldn't protect 16 year olds." Seriously? Is that intentional rage bait?

I think it's more than clear by now that algorithmic feeds are hazardous, at least without significant effort in research and safeguards which nobody seems to be doing. So yeah, I'd say: definitely ban algorithmic feeds for teenagers. Hell, ban them for everyone if you must.

Gating should be done either by ZKP (zero-knowledge proofs, which don't expose any information to any party other than "I'm at least x years old" -- look this up if this is a new concept to you) or device-side by standardizing and streamlining child safety locks.

[–] thatonecoder@lemmy.ca 18 points 8 hours ago (2 children)

No, just go for the jugular and ban such “algorithms” for EVERYONE, no exceptions.

[–] avidamoeba@lemmy.ca 4 points 4 hours ago* (last edited 4 hours ago) (1 children)

Ban American corporate social media altogether? 🍁 Fedecan 🍁 take over with Lemmy, Mastodon, Pixelfed and Friendica?

collapsed inline media

Nevermind, we'll get invaded for that.

[–] thatonecoder@lemmy.ca 2 points 4 hours ago (1 children)
[–] avidamoeba@lemmy.ca 1 points 4 hours ago

I mean I wouldn't say no if it comes down to that and I'd probably sign up to work for the technical effort.

[–] jaselle@lemmy.ca 4 points 8 hours ago

You say "no," as though you are disagreeing with me. But did you notice I said this?

Hell, ban them for everyone if you must.

[–] kbal@fedia.io 4 points 8 hours ago (2 children)

zero-knowledge proofs, which don't expose any information to any party other than "I'm at least x years old"

Not quite. The well-known zkp for age verification used in the obvious way reveals only: 1. "I'm at least x years old" and 2. "my name is y." The name can be some other unique assigned identifier, but the point is that whatever is used it needs to uniquely identify you.

There is no way to tell how old people are across the Internet without relying on unprecedented and shocking intrusions into our privacy.

[–] jaselle@lemmy.ca 2 points 8 hours ago (1 children)

I'm not convinced that ZKP requires an identification number or any such deanonymizing data. If there is a ZKP protocol that implements this that is just one possible implementation.

[–] kbal@fedia.io 3 points 8 hours ago (1 children)

How would you get by without one? If I produce a proof right now that I'm at least 32 years old, how else do you know it's a proof for anyone in particular and I didn't get it from my older brother or some random website that sells them?

[–] jaselle@lemmy.ca 1 points 6 hours ago (1 children)

Well, that same problem exists with many of the proposed verification models, like credit cards (how can you verify this is my credit card?) , photo ID, etc.

Here's my proposal: your browser can send a request to a verification body (could be the government directly, let's say) to respond to the challenge from the website you're accessing, without sending information about which website is asking for the challenge. The verifier sends a cryptographically-signed approval back. The browser forwards this to the website. To prevent comparisons of timing as a deanonymization method, the browser can wait a random period of time before forwarding the request both ways.

[–] kbal@fedia.io 1 points 5 hours ago (1 children)

Every time I've looked at the details of elaborate schemes resembling the one you imagine, I'm always left with a lot of doubts that they're secure or practical. Every time I've looked at the systems that have actually been implemented in reality, I have no doubt that they suck.

[–] jaselle@lemmy.ca 1 points 5 hours ago

That's valid. My preference is for device-side child locks. For instance, a header that says, "I am a child." There is much to improve there still. But failing that, if the winds of politics dictate we must have verification -- why not ZKP?

[–] avidamoeba@lemmy.ca 2 points 7 hours ago* (last edited 7 hours ago)

I read about a cross-signing scheme where diff gov't agencies can cryptographically sign an ID that allows only partial information to be shared with any one service provider. It was done by some institutions in the nordics.

With that said, our government is already trusted with our personal ID information. Nothing stops us from creating public service which can be queried for age, which would only provide an answer after the explicit approval by the person through another channel (e.g. email to sign into gov't portal and approve the age query request). Then require service providers to use it. In fact Equifax already offers such a service without our consent but it costs money to query.

[–] avidamoeba@lemmy.ca 2 points 7 hours ago* (last edited 5 hours ago)

Age-restricting corporate social media isn't "kicking teens offline." That's a funny straw man.

We need age restriction and regulations on moderation and algorithms. The latter alone won't solve the problems social media poses for developing brains. Age restrictions aren't bulletproof and that's alright. It's much easier to stop my child from smoking at the age of 10 when there's a smoming ban in place than when there isn't. I want it to be easier to raise them without developing prepubescent brain rot than not. And I think my neighbours would appreciate bringing up another Canadian that has their marbles intact.

E: Plenty parents outside of the terminally-online circles don't even realize they should restrict social media use at an early age.

E2: Tha fact that the Australian ban doesn't deal with the ID problem is a problem that I definitely would not want us to emulate. A problem in that it does not forbid ID collection by private corporations and it does not provide a privacy-preserving public service for proving age. Besides, Meta already knows the age of most of its users. A reasonable compliance criteria could be established that isn't 100% that would also be good enough, subject to regulatory audits.