Does anyone have a citation for this that wasn't written by Claude? It wouldn't surprise me, but I refuse to look through AI slop to check the accuracy of the report.
(5)(a) "COVERED APPLICATION" MEANS A CONSUMER SOFTWARE APPLICATION THAT IS ACCESSED THROUGH A COVERED APPLICATION STORE AND THAT MAY BE RUN OR DIRECTED BY A USER ON A DEVICE.
(b) "COVERED APPLICATION" DOES NOT INCLUDE:
(I) A SOFTWARE APPLICATION THAT DOES NOT PROCESS USERS' PERSONAL DATA; OR
(II) AN APPLICATION FROM A FREE, PUBLICLY AVAILABLE CODE REPOSITORY.
So if your service is proprietary, but your client is open source, it looks like your're free to go.
As someone that relies on third-party clients to get usable interfaces, if this gets widely adopted it would be great news. It would end the cat-and-mouse game from companies trying to force users onto first-party clients.
But only if the user is not getting the app through an app store but from a "code repository"? I'm not sure if I interpret that correctly, but at first glance it seems confusing and ambiguous.
Does that mean I need to download the Android apk from a git repository? Would a clever lawyer be able to argue that the release section on GitHub is outside the repository and therefore not fulfilling this clause?
Would F-Droid still not be exempt because it is structured like a store and offers pre-built binaries?
But the text says "or," not "and." So by my interpretation if you process user data but are available via "free, public" repo, you're not covered. I presume "free" is defined elsewhere in the text, and that it approximates "open-source."
On the one hand, I'm absolutely against blanket age verification laws like this one, think there are better ways to solve the stated problem, and believe that the current crop of legislation is being pushed by bad actors for nefarious purposes by means of pandering to public mania.
On the other hand, I do appreciate that a possible unintended consequence of the out provided by (5)(b)(I) could be that PII (along with user generated content in general) becomes similarly radioactive to if the US had passed a GDPR equivalent. Either that or it's used as a justification for every single online service to require government ID in order to interact with it "because liability". Unfortunately I assume the latter is somewhat more likely at this point.
Also is it defined precisely what it means to "process users' personal data"?
> there are better ways to solve the stated problem
Call your representatives. There is overwhelming demand for age gating social media (based on, honestly, good evidence). This will be implemented based on who calls in. If the status quo of technical people being hopelessly nihilistic continues, it will be written in the stupidest ways possible.
Can't say I agree. Notice that the proposed legislation isn't specific to social media. Rather it's explicitly advanced in support of Colorado's data privacy laws as they apply to minors.
There's evidence of lots of different issues, a few age related but most not. Adults certainly aren't immune to adversarial algorithms and dark patterns and the practical need for privacy isn't limited to children. It's more that we only seem to be able to achieve broad consensus to add additional regulations where it concerns children.
It's always written in the most midwit way possible, then, once predicted failure happens it's patched up to be slightly better. That's the default assumption for most of the things.
What do you expect? American politics selects for mediocrity. Being a world-class expert on something is a career disadvantage. Most of the electorate wants wants bullshit artists and cartoon characters.
TL;DR We need age verification laws to prevent minors from accessing the addictive stream of toxic sludge rather than outlawing its manufacture and distribution.
How exactly would you do this without, you know, violating the first amendment? Algorithmic feeds are nothing without the content. People get toxic sludge because they signal to the algorithm that they like that.
It’s just the algorithms promoting things I want banned.
You may choose to sign up to see all the toxic sludge you wish, as is our constitutional rights as Americans.
You say “they signal to the algorithm”, but how? How did they see it in the first place to be able to provide that signal? It was suggested to them.
Often because that kind of content is really sticky for the site. Whether because you like it or it outrages you or scares you it’s manipulative in a way that is symbiotic with the platform’s goals.
It provides perverse incentives for creators and companies.
> It’s just the algorithms promoting things I want banned.
And again: the only reason the algorithm promotes things is because that person signaled that they were interested in it. They might've gotten it recommended by a friend, acquaintance, whatever, but the point is that if nobody had recommended anything to them the algorithm would have no data.
And again: how do you propose to get this to survive the first amendment? Algorithms are a form of speech under law.
Presumably by outlawing the types of algorithms used with the legislation carefully limited to a particular context rather than anything being authored by an individual. Right to express oneself preserved, government regulates a harmful product, business as usual.
As far as this specific Colorado legislation goes (which is concerned with the ability to comply with their previously passed data privacy law) I think it's not entirely bad but I have two issues with it.
First, it reverses the problem. Services should be sending an age-appropriateness (or even just general content classification) signal to the device for local processing, not the other way around. If you're going to mandate that OS creators do anything it should be to implement a certain baseline level of (interoperable!) functionality as far as parental controls are concerned.
Second, the entire thing should be predicated on some metric such as MAU or revenue or combination thereof not on the exceedingly vague idea of a "free, publicly available code repository".
> First, it reverses the problem. Services should be sending an age-appropriateness (or even just general content classification) signal to the device for local processing, not the other way around.
OK, so say the device receives a signal that say that an app is not appropriate for children under 13. How would the device find out if the user trying to run the app is under 13?
No, the mania is based on extremely bad/cherry picked evidence. There are at least 6 studies alone (some including meta-analysis) which has found absolutely no link to prove social media is addictive or harmful to children. If anything, they've found the opposite, and one even suggests that calling it addictive might be causing the very problem we're pretending to solve
That wording could be interesting, because it's ambiguous if free is applicable to the repository or the project. Presumably, the latter. This means you could absolutely do source-open but not open-source and still get around it.
Well it says code repository not artifact repository. But it doesn't prohibit obfuscation or transpilation and more generally doesn't appear to specify anything beyond "free and publicly available". I really get the feeling that the people who wrote the law don't have a clear idea of what they're trying to say here and that any court decision is going to be a roll of the dice.
"It's only for porn sites" to "its only for social media" to "its doesn't include open source projects" to "its only when you need an internet connection".
That's how politics works actually. Something has to be done but also not upset X, Y, Z because they will be loud. It's quite okay situation when it happens I think.
Yeah. I think a lot of us just look at computers and operating systems differently than these legislators. But we need to more effectively communicate our needs and side effects of their policies. And elect younger folk sheesh.
I'm actually OK if websites trend toward being endpoints and there's competition for frontends. The unification of the two by site owners has been a net negative for the internet.
So a FOSS app running a device local diffusion model specifically for porn would be free of age checks. From a technical perspective that's not all that different from, say, an ansible playbook or bash script or whatever to download a model from HF and configure a local inference stack yet I feel like it must be an unintended loophole.
As someone working on an open source project in CO, this is a welcome fit of common sense. How do these laws typically work in other jurisdictions, do they block non-conforming sites? Or does it open you up to lawsuits?
Edit: It looks like these laws will be enforced by app stores primarily, because they have more significant liability. I'm guessing they won't take the effort to provide exemptions to jurisdictions with the open source carveout unless it is common.
Of course you do. And farmers like subsidies for corn. That's a general idea for them too. And of course you're going to say the public benefits from open source projects and the farmer will say starving no good. Middle class see, middle class do but think they no do.
We have age verification for many things. The problem now is trust. There is, for obvious reasons, negative trust that this won't ultimately harm people. That it won't be used to harvest more data and invade our digital lives even more. That negative trust is there because we see a constant ability to gather even more information about us, and use it to produce real harm, but no hint at an entity actually fighting back to protect people. If anyone in any government is reading this, you do not gain my trust that big tech will not abuse my information by requiring big tech to collect more of my information, you just loose my trust in the government. Earn my trust back and then, maybe, in some distant future, we can talk about 'but who will think of the children' legislation like this.
What for? That's kind of strange. Maybe if its a critical project, but for random projects that aren't like apache web server, nginx, or Linux Kernel, I don't care, heck I would argue if its a very very small change, and it has been scrutinized I don't care who it came from.
Whoever is behind this needs to be exposed, tarred, and feathered.
Meta's well know campaign was actually to make the app stores (and maybe OSes) responsible for age verification, not apps.
Google and Apple campaigned to make apps responsible for it.
(b) "COVERED APPLICATION" DOES NOT INCLUDE:
(I) A SOFTWARE APPLICATION THAT DOES NOT PROCESS USERS' PERSONAL DATA; OR
(II) AN APPLICATION FROM A FREE, PUBLICLY AVAILABLE CODE REPOSITORY.
As someone that relies on third-party clients to get usable interfaces, if this gets widely adopted it would be great news. It would end the cat-and-mouse game from companies trying to force users onto first-party clients.
Does that mean I need to download the Android apk from a git repository? Would a clever lawyer be able to argue that the release section on GitHub is outside the repository and therefore not fulfilling this clause?
Would F-Droid still not be exempt because it is structured like a store and offers pre-built binaries?
It's also naive to believe that a fraction of open source in a companies pipeline would give them a free pass for everything.
On the other hand, I do appreciate that a possible unintended consequence of the out provided by (5)(b)(I) could be that PII (along with user generated content in general) becomes similarly radioactive to if the US had passed a GDPR equivalent. Either that or it's used as a justification for every single online service to require government ID in order to interact with it "because liability". Unfortunately I assume the latter is somewhat more likely at this point.
Also is it defined precisely what it means to "process users' personal data"?
Call your representatives. There is overwhelming demand for age gating social media (based on, honestly, good evidence). This will be implemented based on who calls in. If the status quo of technical people being hopelessly nihilistic continues, it will be written in the stupidest ways possible.
Can't say I agree. Notice that the proposed legislation isn't specific to social media. Rather it's explicitly advanced in support of Colorado's data privacy laws as they apply to minors.
There's evidence of lots of different issues, a few age related but most not. Adults certainly aren't immune to adversarial algorithms and dark patterns and the practical need for privacy isn't limited to children. It's more that we only seem to be able to achieve broad consensus to add additional regulations where it concerns children.
Nah. Can’t stop the money. Let make brain destroying scams and ad spam legal as long as you’re over 18.
You may choose to sign up to see all the toxic sludge you wish, as is our constitutional rights as Americans.
You say “they signal to the algorithm”, but how? How did they see it in the first place to be able to provide that signal? It was suggested to them.
Often because that kind of content is really sticky for the site. Whether because you like it or it outrages you or scares you it’s manipulative in a way that is symbiotic with the platform’s goals.
It provides perverse incentives for creators and companies.
And again: the only reason the algorithm promotes things is because that person signaled that they were interested in it. They might've gotten it recommended by a friend, acquaintance, whatever, but the point is that if nobody had recommended anything to them the algorithm would have no data.
And again: how do you propose to get this to survive the first amendment? Algorithms are a form of speech under law.
As far as this specific Colorado legislation goes (which is concerned with the ability to comply with their previously passed data privacy law) I think it's not entirely bad but I have two issues with it.
First, it reverses the problem. Services should be sending an age-appropriateness (or even just general content classification) signal to the device for local processing, not the other way around. If you're going to mandate that OS creators do anything it should be to implement a certain baseline level of (interoperable!) functionality as far as parental controls are concerned.
Second, the entire thing should be predicated on some metric such as MAU or revenue or combination thereof not on the exceedingly vague idea of a "free, publicly available code repository".
OK, so say the device receives a signal that say that an app is not appropriate for children under 13. How would the device find out if the user trying to run the app is under 13?
"It's only for porn sites" to "its only for social media" to "its doesn't include open source projects" to "its only when you need an internet connection".
A colleague is hosting a virtual session on these and other similar bills around the world in two days https://maintainermonth.github.com/schedule/2026-05-22-age-a...
Or, now slightly out of date, read https://github.blog/news-insights/policy-news-and-insights/w... Added: I had not scrolled far enough on the front page, https://news.ycombinator.com/item?id=48214215 is on this blog.
Names matter. We saw ChatControl 1.0 get defeated, it probably didn't hurt that the name implied censorship.
Edit: It looks like these laws will be enforced by app stores primarily, because they have more significant liability. I'm guessing they won't take the effort to provide exemptions to jurisdictions with the open source carveout unless it is common.
Annoyed by the age gating, or feel it to be commercially burdensome? Open your source, and poof, no more mandate!
Just trying to build and maintain a cool thing, and share it with the world? Never mind the compliance burden.
(https://www.therage.co/tag/tornado-cash/)
I feel like age verification is important online - a copy of the real world. Check my ID before I go in the pub.
It feels like it's jumped all the way to positive-ID. Not just "of age" but become you are "First Last".
It's possible (right?) to assert age and is-human attributes w/o knowing which specific human at what specific age I am online?