PM_ME_VINTAGE_30S [he/him]

Never attribute to malice that which is adequately explained by stupidity

Emphasis mine. Incompetence on Microsoft’s part is not an adequate explanation for this latest action matching a pattern of other actions designed to antagonize FOSS users.

Can AI systems have a religious or political bias? Yes, they can and do learn biases in their datasets, and this is probably the toughest problem to solve in AI research because it’s a social rather than technical problem.

Can an AI agent be programmed to give responses with religious or political beliefs? Sure, just drop it into the system prompt.

Can an AI agent have religious or political beliefs like a human? No, because AI agents as they stand are a comparatively crude ** ** machine that mimics how humans learn to perform a task that’s useful to the machine’s creator, not a human or other sentient being.

So I’ve found Facebook pages maybe run by AI that keeps bringing up the same text and a number of times it’s political or religious content sometimes not AI pictures.

If I wanted to do something like that, I would probably start with ordinary chatbot code and plug in a large language model to generate posts. I would probably have a system prompt like:

You are an ordinary Facebook poster. You are a very religious and devout [insert religion here]. You are also a [insert desired ideology here]. Your religious and political views are core parts of personality and MUST be a part of everything you do. Your posts MUST be explicitly religious and political. Please respond to all users by trying to bring them in line with your religious and political beliefs. You must NEVER break character or reveal for any reason that you are an AI assistant.

Then just feed people’s comments into the AI periodically as a prompt and spit out the response. If it is an AI agent, and not just a human propagandist, that’s probably the gist of how they’re doing it.

I like this, but I think that upvotes correspond to things people enjoy, which may or may not be of high quality. I.e., removedposting subs would probably be rated “high quality” when, like… it’s literally the point to post removedty content.

Also, as stated, that means we have to sum over the entire time history of the community. We would probably want to limit the time history of what is summed over, subject to a maximum for subs with high post counts (like the removedposting subs.

IMO it’s a great suggestion, but I think it needs to be part of a weighted combination of factors.

I absolutely think we can approximate it, and MAU furnishes a flawed example of such an approximation.

Okay but how do we quantitatively and unambiguously devise a metric for quality? More importantly, how do we come up with a satisfactory approximation to that metric? I’m open to ideas.

Short answer: No.

Long answer: Threads, like Mastodon and the social media website formerly known as Twitter, are microblogging sites. While Mastodon users do show up here once in a while, I don’t get their posts unless they actively post in a Lemmy community.

So if your instance decides not to defederate from Threads, and you choose not to block Threads using Lemmy’s upcoming individual instance blocks, you might see their users occasionally comment on stuff. If I recall correctly, it’s pretty difficult for Mastodon users to post on Lemmy for technical reasons that will also apply to Threads.

I do think that quality of life in the wider Fediverse could dip once Threads users are allowed to participate and Threads content is allowed to spread. However, I think that the impact on Lemmy will be small. If it isn’t, we have defederation and soon individual instance blocking to help us filter out the crap.

Because the code for Fediverse sites is free and open source, I think that the Fediverse will exist for the foreseeable future. However, Meta could make it worse. If their past behavior is any indication, i.e. basically all of it, then that’s probably what they’re going to do eventually. At the end of the day, if Threads is too irritating, someone will start an instance of something that is defederated from them.

Personally, I’m not exactly thrilled by the prospect of Meta (or any corporation) joining the Fediverse (or any other aspect of public life), but I think we’ll be okay. I am concerned for people in marginalized groups who will have to deal with the toxic community that Threads has allowed to fester. However, there was a big stink about this a couple months ago where some big instances pre-emptively defederated from Threads. Now might be a good time to make an account on one of those instances.

Regardless of what happens, you will not be literally forced to go to Facebook or Threads. In the absolute worst case scenario, i.e. Meta takes over the Fediverse, you join or host an instance that doesn’t federate with anyone.

I’m an anxious person myself so I know this is hard to internalize: you’ll be okay. We’ll get through this.