#Ethical Data Use

[manmishra]

🚀 Dive into the future of data storytelling! Discover how AI and innovative tech are transforming the way we communicate insights and engage audiences. Explore the essential role of human creativity in this evolving landscape and learn how to leverage these tools for impactful narratives. Read more about it in our latest article! 🌐📊 #DataStorytelling #AI #Innovation #MarketingStrategy

[ganondoodle]

the company i work for decided that its switching from the german formal "You"(Sie) to the informal "you" (Du) in all of our websites so now we have to scour the entire database to change it and i quite frankly hate that, not just bc the unecessary extra work but especially bc its such a weird and unecessary change

i bet its bc everything here is getting englishfied (both literally and culturally it feels like, when my new boss talks its half in english bc every second german word is just replaced by an english one despite there being perfectly fine words for it in german too, its so annoying) and bc they want to sound more personal in hopes of getting more clients bc 'company is your fwiend uwu!!', i know this here is the amercian tm site so you wouldnt understand really but i do not want to be greeted with 'du' by companies, no, thats too personal, you dont know me and im not giving you my data, stay away!!

i guess thats how i would describe it .. the formal you is like a polite distance, like someone you dont know staying outside your personal space, but when its the informal 'you' it feels invasive unless i told you you can call me that, and that goes double for companies

maybe its a small thing that doesnt seem important but i cant stand it, im just a little part time worker doing data work so i got no say in it but the companies founder also announced hes giving his post to his kids some time ago so ...... since then theres been alot of changes and new projects that solely aim to imitate whats popular and whats done by other companies, despite ours being one that is, or used to be, intentionally different, like, that was the POINT, but i guess chasing trends is just too appealing for CEOs

[darkwood-sleddog]

you know i think it's really funny (derogatory) that tumblr seemingly has the ability to create popups for ad-free, crabs, anything that will get them money, but when it comes to using popups to guide new users around the website layout they're just like "uww what can we POSSIBLY do?? all these new people only know twitter and they couldn't possibly read a popup explaining how to use the site!! let's just change everything visually appealing about tumblr to cater to our inability to make popups for helpful reasons uwu!!".

like you know what i want? If you're testing a change on me I want a pop up that says so. I don't want to have to rely on the @changes blog to see that. If i'm testing a change I should be given a feedback form that is specifically about that change. I want the results of the feedback forms for changes to be publicly viewable to the userbase. you could learn so much about how this website actually functions for users through that, but no. we can only make popups when it fiscally benefits us! we can't use polls to get a sense of what the userbase is feeling! we can't publicize what % of feedback about a change was positive or negative. we couldn't...possibly...

[discolesbo]

Fandom people if you go to Threads you're some dumb motherfuckers

[kaurwreck]

please send help, my boss is using ai to generate pop songs about the end of globalization.

[i-will-not-be-caged]

Got a request from a coworker today that was basically like “I have this narrative I want to tell; can you give me data that will support it?” and WOW did every fiber of my being cringe when asked to do data analysis in exactly the way you’re not supposed to

[powerfulkicks]

man i hate the current state of ai so much bc it's totally poisoned the well for when actual good and valuable uses of ai are developed ethically and sustainably....

like ai sucks so bad bc of capitalism. they want to sell you on a product that does EVERYTHING. this requires huge amounts of data, so they just scrape the internet and feed everything in without checking if it's accurate or biased or anything. bc it's cheaper and easier.

ai COULD be created with more limited data sets to solve smaller, more specific problems. it would be more useful than trying to shove the entire internet into a LLM and then trying to sell it as a multi tool that can do anything you want kinda poorly.

even in a post-capitalist world there are applications for ai. for example: resource management. data about how many resources specific areas typically use could be collected and fed into a model to predict how many resources should be allocated to a particular area.

this is something that humans would need to be doing and something that we already do, but creating a model based on the data would reduce the amount of time humans need to spend on this important task and reduce the amount of human error.

but bc ai is so shitty now anyone who just announces "hey we created an ai to do this!" will be immediately met with distrust and anger, so any ai model that could potentially be helpful will have an uphill battle bc the ecosystem has just been ruined by all the bullshit chatgpt is selling

[insightfultake]

Why Did India’s Finance Ministry Restrict the Use of AI Tools in Offices? A Closer Look at the Decision

In a significant move, India’s Finance Ministry recently issued an advisory restricting the use of artificial intelligence (AI) tools, such as ChatGPT, Bard, and other generative AI platforms, in government offices. This decision has sparked widespread debate, with many questioning the rationale behind it. Why would a government, in an era of rapid technological advancement, curb the use of tools that promise efficiency and innovation? Let’s delve into the logic and reasoning behind this decision, including the geopolitical implications and the growing global AI race, particularly with China. Read more

[size-two-shrimp]

I loooooove trying to find a picture of a certain fish's fins because they're so hard to see and multiple images are both AI generated and completely fucking wrong!

[professor-rye]

I feel like, with the uproar over Nanowrimo right now, we have an opportunity to really push back at shitty AI, but I feel like we also need to be smart about it.

Just saying "Generative AI is bad! Fuck you!" is not going to make a huge dent in shitty ai practices, because they'll just dismiss us out of hand. But if we ask the really hard hitting questions, then we might be able to start making some level of progress.

Mozilla is actually doing a ton of good work towards this very goal.

They've been working to try to shift industry goals towards more transparent, conscientious, and sustainable practices, and I think their approach has a lot of promise.

AI is not inherently bad or harmful (hells, even generative AI isn't. It's just a tool, thus neutral at its core), but harmful practices and a lack of transparency make it to where we can not fucking trust them, at least in their current iterations.

But the cat is out of the fucking bag, and its not going back in even if we do point out all the harm. Too many people like the idea of making their lives easier, and you can't deny the overwhelming potential that AI offers.

But that doesn't mean we have to tolerate the harm it currently causes.

[alxfsl]

Apparently Matt is not trustworthy as I thought

[midnightactual]

to talk about AI in art more specifically, I think in so far as AI has any legitimate use in art, it's exclusively as an artistic aid. if you generate an AI image of a character, then draw on top of it using it as a loose reference (I'm not talking about tracing here), that's not really any different than using any other photo reference or inspiration. if you feed all of a character's dialogue to an AI, then have it crosscheck versus dialogue you wrote for them and make adjustments to your writing (I'm not talking about copying and pasting here), that's not really any different than having someone beta-read for voice or doing it manually yourself. it can be an effective tool for doing very specific and technical things faster and thus developing a better personal artistic intuition for them. it's acceptable as training wheels, like spellcheck but much more sophisticated

what is broadly unacceptable is to just let AI generate something and release it into the wild as-is or edited. AI making your life easier or giving you more confidence in making your art is fine. using AI to substitute for and try and shortcut the artistic process is not

[hychlorions]

i've had it so i am starting to try and convince my friends to drop character ai

[cosmereplay]

Well, I lasted less than a month of locking my ao3 fics to registered users. It didn't have any impact at all on the problem of data scraping, and having little to no interactions on ao3 has been impacting my motivation to write.

I really want to write, I enjoy writing on its own, but I gotta say it's a real boost when other people are reading and enjoying my works.

So, I've unlocked them to the public again. You can check out my Stormlight fanfic here (beware the broad range of ratings and topics!).

Instead, I'll use this opportunity to tell those who live in the States that a law firm called Clarkson is preparing a class action lawsuit against OpenAI over their data scraping practices. You can check out the law firm's website here, and join and voice how you've been affected.

[byakuyasdarling]

I think a good thing about being away from social media though is just caring so much less about all the internet argument stuff. It’s so much less stressful just focusing on me and my health and the people close to me.

Especially with AI stuff. Of course I don’t agree with it scraping from artists but I love when artists reclaim it as a tool and I think it should be used as such. You can’t stop a program from existing, it’s useless. But you can make guidelines to ensure it’s uses are ethical and practical — basically to make jobs easier and not over-work artists, not to replace them.

I think there’s so much to still work-out in that regard, obviously.

Another thing that used to stress me out was those “press 3 buttons to save my pet” videos. I always try to do the copy link think and get interactions up but it started triggering me my anxiety which wasn’t good. Have you guys experienced that, and how did you deal with it? /gen

[wrenfea]

The burden of keeping training data free of copyrighted data-poisoned works should be on the company using it, not on the people just trying to protect their works.