Should social network algorithms be publicly available?

Recommendation providers should be decoupled from content providers. Not unlike movie studios and movie theaters. Open-sourcing current algos won't help, they are flawed, but their flaw isn't really in the lack of openness. It is in the lack of competition. If you make the best recommendation algorithm in the world that is miles ahead of everything in existence, you wouldn't be able to make a product out of it, because you need the whole package in order to compete, so your only options are either to try to caught up on features to companies that have vast profitable businesses or sell your creation to said companies. And, by the way, voting on a changes to a single algorithm that makes choices for everyone won't help since a lot of people are going to be pissed of by any decision that is made, which, in turn, means that everyone is going to be pissed off by that algorithm since virtually everyone is going to see a decision they don't like being passed(and so we are back to square one).

Not just publicly available. They should be reproducible.

I image a scenario of a journalist who wanted to investigate whether some post on a social network's wall was pushed by a group of certain interests or maybe analyze the trends and understand who's behind them. There should be a way to receive a set of inputs and a reproducible way to get the same output.

It could start with a user interface where a user could click on a post and they would see the full context of what has led to this post being shown to them at a given moment.

I think it would help eliminate the current trend of shoving whatever machine learning models deem the most addictive and would make the web more transparent and user friendly because then the developers would have to optimize differently.

So to summarize, I think the algorithms AND datasets should be public. In addition to that, even if they were not public, their inputs and outputs should be reproducible and accessible.

Give me a choice of algorithm from a list of algorithms. Promote your choice by making it the default on your platform.

FOSS social network (like Mastodon) with FOSS algorithm for content recommendation would be the ideal place to be: therefore you would have competition on all aspects: clients, algorithms, and all in the open.

Yes. Honestly the “algorithm” should just be your followers’ latest status updates chronologically - but there’s no accidental or hyped up ad clicks in that.

Personally I think it would be interesting if each state/province or whatever would host it's own people's speech. The public ought own an interoperating, internetworked systems of public speech.

I dont think Elon's complaint mirrors yours. I dont think his offence is chiefly targetted at algorithms. The need to sign in, evermore people being banned, the ongoing elaboration of what is allowed not allowed- these are real barriers. Personally algorithmic curation has not denied me access to my friends- it simply mixes in other things.

I'm fine with whatever algorithms social networks want to use. What I'm not fine with is them being forced on me to show me things I didn't subscribe to. I just want to know what my friends are up to.

In my view, social networking plays an enormous part in today's social ills. I'd like to go back to a simpler time when we were in greater control over the media we consumed -- not by forcing the algorithm to change, but just by opting out of it altogether.

You vote with your feet and don’t use social media. There’s no “free speech principle.” Private companies can allow or censor whatever they want to on their own platforms.

Just the algorithm being publicly available won’t help much, unless all the data that is input to the algorithm is also publicly available.

I think social media and recommendation algoritms will get regulated. Since social media algorithms might be built to cause addiction. Same principle as tobaco industry, eventually it will be regulated since it affects health.

These are companies with trade secrets and algorithms are trade secrets. You don’t disclose such as that is your competitive advantage.

any algorithm will be gamed so it doesnt make a lot of sense. Why do we have algorithms again? My feed is chronological, which forces me to only include the right amount of people, and the right people. When twitter randomly resets it to algorithmic, it is noticed, and it's always a downgrade

Sure, why not. However, will they be? No.

I will say no. The algorithm isn't the problem at all. There is 1 thing and 1 thing alone that is the problem.

Twitter and others exist entirely because of Section 230. That as a platform they will not be held liable for the content they host.

But there's a caveat. They may only censor content within very specific categories which most people would agree with.

>No provider or user of an interactive computer service shall be held liable on account of-

>(A) any action voluntarily taken in good faith to restrict access to or availability of material that the provider or user considers to be obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable, whether or not such material is constitutionally protected; or

That right there is the extent twitter may censor. They cannot go beyond this but they can go less. You are not obligated to remove all lewd things from their website. Twitter clearly does not given how much porn and gore they host. Twitter actively allows death threats. Good faith is super important in this rule.

So lets roll this up into what exactly is the problem in the context of current events. Rachel Levine and mass banning of conservatives.

Rachel Levine was unimpressive achieving nothing of note. She got appointed to a 4 star admiral position solely because of being trans. Then got labelled as woman of the year? Where is the achievement? Just immediately right to 4 star general and woman of the year?

That's not even the controversy. The controversy is that Twitter is banning people for saying "Richard levine spent 54 years of his life as a man" or "men aren't women" not even naming anyone or even linking trans issue. It is a fact that men aren't women. That's why they are separate words.

The conservatives literally wrote what is factual and got banned. They intentionally crafted it this way to show the bad faith.

Now let's roll back to section 230. Twitter is acting in bad faith to restrict specific viewpoints. In so doing creating a protected class of people. Not unlike how twitter actively allows Putin or various other tinpot dictators but doesn't allow trump? It's all bad faith by twitter.

The problem is obvious. Twitter is in clear violation of section 230 but the government refuses to enforce this rule. So Twitter gets all of the liability protections and none of the requirements of free speech.

Now go back to conservatives who just got disenfranchised. Do you think they are going to take the ban and say, you know what maybe I was wrong about trans people. No. They are going to be radicalized more.

Twitter is creating hatred for trans people. As a trans person, twitter is one of the key players in why trans people are hated, and worse the coming consequences for this will not be good for us trans people.

Now imagine twitter got a legal letter saying section 230 good faith will now be enforced. That a single ideological ban like "men aren't women" will be an immediate suspension of their section 230 protection and they will be held as a publisher.

This immediately solves the problem. Twitter doesn't exist without section 230, they would be forced to immediately solve this huge problem.