[diacritical]

In theory I'm OK (kinda) with regulating the "overall design" somehow, but I don't see how it's going to work. Forced interoperability is a (very?) good idea, as it's really general, but it also doesn't address directly what the article and most comments talk about - the rage bait. I just can't imagine regulations (or "laws" or whatever the correct term is) that deal specifically with the algos that push rage bait that can't be later abused, if passed, to deal with other unpopular speech. And it seems like people want some laws to directly deal with that - the bad types of speech or algos themselves.

To clarify, I use "rage bait" as an example phrase, but it includes algos that only promote engagement at any cost and other things that aren't outright dangerous, but we think are dangerous. Not, like I said, CSAM or yelling FIRE or telling people to kill themselves.

[fc417fc802]

Interoperability sidesteps the issue by giving users the choice of which algorithm (or algorithm provider) to use. The majority might or might not agree with that approach - for example obviously tobacco has not been left purely to the individual's judgment in the west.

Agreed, you can't regulate speech in a targeted manner while also not doing so. You're forced to find some common aspect much more general than "rage bait". Perhaps prohibiting the targeting of certain metrics? Or even prohibiting their collection in the first place.

[diacritical]

> You're forced to find some common aspect much more general than "rage bait". Perhaps prohibiting the targeting of certain metrics? Or even prohibiting their collection in the first place.

Can you elaborate, give some ideas, examples, etc.? What metrics? How can you define them in a consistent, safe way?

[fc417fc802]

We're talking generalized metrics. I have no idea which ones - I wasn't claiming to have solved the problem. The point is that if you can identify a general characteristic that is being used in a way which disproportionately contributes to a particular outcome then you can filter on that.

Estimated user age is an example of a metric largely unrelated to concerns regarding free speech. I doubt it has much relevance to the problem we're taking about here but hopefully you can imagine that prohibiting the targeting of ads or the curation of an algorithmic feed based on that metric would not be expected to unduly disadvantage any particular sort of speech.

[diacritical]

> The point is that if you can identify a general characteristic that is being used in a way which disproportionately contributes to a particular outcome then you can filter on that.

In a non-adversarial political context where we trust the government and the future ones, sure, but I think without any strong guardrails, we could enact a law that's good today, but will be exploited in the future.

For targeting minors with any kind of political speech - I'd love it if it wasn't legal. But that brings its own can of worms. There's enough discussion on HN on the implications of age verification, whether on how it's done technically (privacy-preserving or not (ZKP or just shady 3rd parties); FOSS or not; on the ISP, OS or app level, etc.) and whether the mere precedent could trigger additional issues down the road.

Anyway, I'd love a society where everything is perfect, but I'm afraid of what might actually happen. With a benevolent god as a permanent ruler, I'd be happy with 100% prosecution rate against all kinds of littering, hate speech and whatnot, but in reality random crimes are easier to evade than a law passed down by a malevolent government, so I'm strongly against any kind of overreach. (Because the law tomorrow could be one we must evade if we want to resist an unethical government). Someone will likely chime in with "but complete and massive overreach has never happened so far", to which I'd reply - we're close to the point where technology will let the ones in power grab that power absolutely and forever if we them grab too much in the beginning.

[fc417fc802]

> In a non-adversarial political context

An oxymoron.

> where we trust the government and the future ones

Has never and will never exist.

> we could enact a law that's good today, but will be exploited in the future.

Sure but that's how pretty much all legislation works.

> I'd love it if it wasn't legal. But that brings its own can of worms.

It's probably fine as long as you include the clause "knowingly and intentionally". That doesn't imply age verification or anything else, merely that you act on information that you have and are aware of (and that you not intentionally design systems to work around that).

Also note that I never said anything about underage users. My example was targeting based on estimated user age. So in that example the age is estimated and it is illegal to target anything based on the value. (Of course to avoid a very silly loophole you'd also need to disallow targeting based on verified age as well.)