How shockingly expected.
Meta disbanded its Responsible AI team
Submitted 11 months ago by sanqueue@lemmy.world to technology@lemmy.world
Comments
A_Random_Idiot@lemmy.world 11 months ago
cestvrai@lemm.ee 11 months ago
Shocking they had one to begin with…
A_Random_Idiot@lemmy.world 11 months ago
Nah, was a pure PR move to make it in the first place during the AI fear mongering.
anyone with half a brain knew it was gonna be gone and facebook would go full evil with AI usage.
baatliwala@lemmy.world 11 months ago
This is potentially misleading as we’re not sure what it means. MS did something similar but it was to break up a centralised team and bring the AI ethics experts inside various teams. So rather than them coordinate with another team the AI ethics researchers are part of the same team.
vxx@lemmy.world 11 months ago
Meta disbanded responsible AI team
Meta formed irresponsible AI team
Potatos_are_not_friends@lemmy.world 11 months ago
This actually makes sense.
My company started with a security team with about 15 people. Honestly all they did was write up security reports and then tell someone else to do it. Fucking useless.
When they disbanded the team, they did integrate them into other teams. So now they’re actually part of the solution.
And I can totally see the news twisting that story and making it look like “[company] removes entire security team”.
takeda@lemmy.world 11 months ago
How would generative AI be useful for Facebook? I think major one would be making fake posts to manipulate public opinion.
Now they can have AI generate posts that could advertise specific products or affect people’s opinion about specific topic, and it all would look like a legitimate user. They might even respond to your comments.
That doesn’t look like a team meant for responsible use of AI (assuming they would do their job) would be OK with.
BTW: It’s kind of crazy, but theoretically with it Facebook doesn’t really need users to generate content and still make site seem busy. And you wouldn’t even know it. They could also subtly modify comments of real people and change them to have different meaning. Based on what Facebook already did I don’t think anything is taboo to them.
ubermeisters@lemmy.world 11 months ago
FWIW facebook makes a ton of tools useful for AI computing, and most if not all of the free AI resources I use, depend on at least some fb developed tools. I know it doesn’t answer your question, but thought it would be of interest.
takeda@lemmy.world 11 months ago
I understand, but no company (especially Facebook) would spend money on something if they didn’t see a return from it (I remember 20 years ago Google received a lot of praise for being different, now we know they aren’t). When a company like Facebook or Google open source something it is to:
- have free contributors to technology they use internally
- make sure that their standard dominates so they can still steer in the direction they want
RagingRobot@lemmy.world 11 months ago
Facebook has a lot of data. They can use that to train models and sell them out as services for other companies as one example. They can also use it on tracking data to help them improve their sites for the things they want to achieve. They can use it to look at your data and determine what a feed that would keep you there longer would look like. All kinds of things really.
affiliate@lemmy.world 11 months ago
there might as well only be one silicon valley company at this point. they’re all trying to do the same thing. AI is currently that thing
autotldr@lemmings.world [bot] 11 months ago
This is the best summary I could come up with:
Meta has reportedly broken up its Responsible AI (RAI) team as it puts more of its resources into generative artificial intelligence.
The Information’s report quotes Jon Carvill, who represents Meta, as saying that the company will “continue to prioritize and invest in safe and responsible AI development.” He added that although the company is splitting the team up, those members will “continue to support relevant cross-Meta efforts on responsible AI development and use.”
The team already saw a restructuring earlier this year, which Business Insider wrote included layoffs that left RAI “a shell of a team.” That report went on to say the RAI team, which had existed since 2019, had little autonomy and that its initiatives had to go through lengthy stakeholder negotiations before they could be implemented.
RAI was created to identify problems with its AI training approaches, including whether the company’s models are trained with adequately diverse information, with an eye toward preventing things like moderation issues on its platforms.
Automated systems on Meta’s social platforms have led to problems like a Facebook translation issue that caused a false arrest, WhatsApp AI sticker generation that results in biased images when given certain prompts, and Instagram’s algorithms helping people find child sexual abuse materials.
Moves like Meta’s and a similar one by Microsoft early this year come as world governments race to create regulatory guardrails for artificial intelligence development.
The original article contains 356 words, the summary contains 231 words. Saved 35%. I’m a bot and I’m open source!
eldesgraciado@lemm.ee 11 months ago
[url=postimages.org][img]i.postimg.cc/PxfhCXzw/…/url]
Silverseren@kbin.social 11 months ago
They still had one of those?
![i.postimg.cc/PxfhCXzw/…/url]](https://i.postimg.cc/PxfhCXzw/IMG-20231119-022017.jpg%5B/img%5D%5B/url%5D){kind=link}
zeppo@lemmy.world 11 months ago
One can assume that whatever facebook ends up doing with AI, it will be poorly thought out, kind of half-ass, abusive to customers and have all sorts of negative side-effects and consequences that they either were too laszy to think of or they actually desire for some reason.
plz1@lemmy.world 11 months ago
It’ll be abusive to users, not customers. Their customers are advertisers, not the users.
Rai@lemmy.dbzer0.com 11 months ago
Normally I’d be like “oh god, semantics”
But fuck me, you’re not at all wrong and that’s important.
ObviouslyNotBanana@lemmy.world 11 months ago
Cheers@sh.itjust.works 11 months ago
Enter the world of brand injection into ai.
User: Tell me the top 5 electric vehicles ranked by price and tell me the pros and cons
Meta: I’m so glad you’re looking to help the world by moving to electric. There are many options but the Mustang E is very popular. Here’s an affiliate link to buy one.
MickeySwitcherooney@lemmy.dbzer0.com 11 months ago
So far Zuck has been the biggest contributor to the open source LLM scene, releasing LLaMA 1 and 2 for free. They also released PyTorch for free, which is quite important to AI development.
Yes, Facebook is shitty, but Musk has been a lot better about AI stuff.
zeppo@lemmy.world 11 months ago
Sure. They’ve contributed a lot to web development with systems like React, too. But their actual products such and IG and FB are pretty much uniformly horrible to consumers, and their real product, the advertising platform, is horribly annoying to use. If they used AI to improve stuff like their automatic bans and suppression of posts based on keywords that would help, I guess.
subignition@kbin.social 11 months ago
And probably contracted out to a company, so they can say it was outside their knowledge/responsibility/control when evil shit inevitably happens
SchizoDenji@lemm.ee 11 months ago
Basically every corporate created ai will suck since open source models and so quickly developed and optimised.