> Also I would like an example of something a social media company does that you wouldn't be able to get approval to do on animals.

One possible example is the emotion manipulation study Facebook did over a decade ago[0]. I don't know how you would perform an experiment like this on animals, but Facebook has demonstrated a desire to understand all the different ways its platform can be used to alter user behavior and emotions.

0: https://www.npr.org/sections/alltechconsidered/2014/06/30/32...

> What counts as research? If I make UI changes, I guess it's ok to roll it out to everyone, because that's not an experiment, but if I roll it out to 1%, then that's research?

I think this is a good example of how disconnected and abstract the conversations about social media have become. There's a common theme in these HN threads where everything social media companies do is talked about like some evil foreign concept, but if any of us were to do basic A/B testing on a website then that's understandable.

Likewise, the dissonance of calling for heavy regulations on social media sites or restrictions on freedom of speech is ironic given that Hacker News fits the definition of a social media site with an algorithmic feed. There's a deep otherness ascribed to what's called social media and what gets a pass.

It gets really weird in the threads demanding ID verification for social media websites. I occasionally jump into those threads and ask those people if they'd be willing to submit to ID verification to use Hacker News and it turns into mental gymnastics to claim that Hacker News (and any other social platforms they use like Discord or IRC) would be exempt under their ideal laws. Only platforms other people use would be impacted by all of these restrictions and regulations.

Well, the definition is simply "the systematic investigation into and study of materials and sources in order to establish facts and reach new conclusions"

However, I'd like to narrow this to "The systematic investigation into, and manipulation of, a system in order to map variables that can be exploited for material gain"

The reason this is important is because it's important someone can tinker with their website and test out new aesthetics etc.

As for an example of something you'd be hard pressed to get approval for: Body Image & Social Defeat (The Instagram Study). This would be extremely unlikely to pass. Given the research goal would translate to "we wish to see 'how much bullying makes them stay in the cage longer'" then it would be rejected as gratuitous cruelty.

I think by now roughly half of us grew up in a world where global reach has been simply taken for granted. I don't think it's particularly onerous to say that there should be some oversight on what a business can and can't do in the context where that business is relying on public infrastructure and can affect the whole-ass world, personally

Are you being serious right now or just engaging in "asking questions" to suppress others thoughts? Why are these types of comments so common on this site? No obviously we aren't in fact talking about making basic code changes, but maybe if those changes are being consistently done that clearly show users getting more depressed or alienated it should be questioned more and finally regulated.

Fun fact, the last data privacy law the US passed was about video stores not sharing your rentals. Maybe it's time we start passing more, after all it's not like these companies HAVE to conduct business this way.

It's all completely arbitrary, there's no reason why social media companies can't be legally compelled to divest from all user PII and be forced to go to regulated third party companies for such information. Or force social media companies to allow export of data or forcing them to follow consistent standards so competitors can easily enter the realm and users can easily follow too.

You can go for the throat and say that social media companies can't own an advertising platform either.

Before you go all "oh no the government should help the business magnates more, not the users." I suggest you study how monopolies existed in the 19th century because they look no different than the corporate structure of any big tech company, and see how government finally regulated those bloodsuckers back then.

> What counts as research?

You might be aware of this, but most big tech companies (i.e. the ones with massive user counts) don't just let you roll out UI changes to everyone, because they know that this has a downstream impact on users. So they often A/B test those things, which is literally an experiment: you randomize who sees what, measure outcomes, and ship whatever wins. There are many data scientists employed in industry to set up and analyze experiments like this.

Also, it seems clear that this not harmless research. Everyone is aware of the effect social media has on our mental health (see the under-16 social media ban in Australia). Facebook definitely knows this, e.g. 2014 there was a big controversy over their News Feed “emotional contagion” study, where they altered what content people saw to measure changes in sentiment, without meaningful informed consent [1][2].

> Also I would like an example of something a social media company does that you wouldn't be able to get approval to do on animals. That claim sounds ridiculous.

This misses the main point: the issue is that for these experiments (and they are experiments) there is often no independent approval mechanism in the first place. Facebook, after receiving backlash, does have privacy/integrity/safety teams now which review these experiments, they are far from being independent third parties.

[1] [https://www.theatlantic.com/technology/archive/2014/06/every...](https://www.theatlantic.com/technology/archive/2014/06/every...) [2] [https://www.sciencenews.org/blog/scicurious/main-result-face...](https://www.sciencenews.org/blog/scicurious/main-result-face...)

Nice example there to trivialize and confues the issue but yea if your hypothetical store redecorating has a public health impact on a large scale then you should need approval.