deleted by creator
Everything that is exposed to consumer information will be this way soon. It’s an eventuality at this point, not a maybe.
Only thing I heard about zoom is their back to office?
They changed their TOS to allow themselves to license everyone’s videos for A.I. training (or anything else). One of the execs tried to say they weren’t doing that but unless they change their TOS, they can and no doubt will.
For some people, that’s a personal privacy issue but for people who have Zoom calls about, for instance, health records, it makes Zoom illegal. And even if it’s not illegal, companies use video calls for discussing proprietary information they don’t want to be potentially licensed to competitors.
Wow, wtf…it seems like every big tech company is going through “enshittification”. Is there an open-source alternative for Zoom that is hopefully more privacy-focused?
I haven’t used it but the !selfhosted community has mentioned Jitsy Meet before.
Just discovered it too from !privacy. Reading about it now, it’s FOSS which is always good. Here’s the link: https://meet.jit.si/
There’s element call as well. (https://call.element.io)
Apparently they did:
As of Monday afternoon, [section 10.4] has a new paragraph in bold below it: "Notwithstanding the above, Zoom will not use audio, video or chat Customer Content to train our artificial intelligence models without your consent.”
How exactly they obtain customer “consent” isn’t disclosed.
I’m not so worried about Zoom adding fancy autocomplete (“training our models”) as I am with them licensing it out. This is what section 10.4 says before the caveat:
You agree to grant and hereby grant Zoom a perpetual, worldwide, non-exclusive, royalty-free, sublicensable, and transferable license and all other rights required or necessary to redistribute, publish, import, access, use, store, transmit, review, disclose, preserve, extract, modify, reproduce, share, use, display, copy, distribute, translate, transcribe, create derivative works, and process Customer Content and to perform all acts with respect to the Customer Content, including AI and ML training and testing.
I don’t think that extra caveat even addresses licensing meeting content to third parties for training A.I.
A bit later in the article also addresses this:
That consent, [Zoom Chief Product Officer Smita] Hashim closed, still won’t allow third parties to train an AI off your calls: “And even if you chose to share your data, it will not be used for training of any third-party models.”
However, glancing through the ToS I don’t see where Zoom prohibits third-party AI training, only prohibiting training their own models. On the other hand, data for training LLMs is apparently the modern gold-rush and it’s feasible that Zoom wouldn’t want that data to be accessed by any potential competitors.
you agreed to the TOS, therefore they have your consent.
Does software exists that encrypts video and audio data on one end, and requires a key to decrypt on the other end? Anyone looking at the feed without keys would be seeing garbage.
Zoom already does 256bit AES end-to-end encryption. From what I understand it isn’t the live calls but the files and recorded calls you save on their servers after are what they would use for AI training.
Zoom already updated their TOS a few hours ago to supposedly address the issue. https://gizmodo.com/zoom-ai-privacy-policy-train-on-your-data-1850712655
Update, August 7, 5:06 p.m.: After this story was published, Zoom issued an update to its Terms of Service. The article has been updated to reflect the change.
Oh,forgot to mention there is support for sharing screen and separate window on desktop. Unsure about mobile clients.
I’m not sure about the technicals, but there are some services that are HIPAA compliant, which I assume means something similar to the end-to-end encryption you’re describing. WebeX is the one I know one of my local hospital systems uses.
Zoom and MS Teams both seem to have E2E encryption for 1-on-1 calls. They own the code, tho, so whether they really cannot decrypt the stuff is a matter of trust.
Yes. End to end encryption between two parties is pretty trivial to implement. Video and audio is no exception.
Yes, E2EE video call software exists, but then you’d be using something other than Zoom.
Theoretically you could have everyone in a call use a plugin that “added” E2EE to Zoom, but I’m not certain such a plugin exists - and even if it did, ensuring everyone you communicate with Zoom uses it would be enough of a barrier that it’d be simpler to just use something other than Zoom that has E2EE baked in.
Holy shit how did I not hear about this but the back to the office thing was everywhere? Fuck that, no. No that is not ok. Fuck. FUCK why is everything going to this shit?
They also gave an opt-in to allow all your data to be used to train AI models. You auto opt-in if you use any of their AI assisted services
What did zoom change? Other than their workers being forced back to offices
Hot take
The internet was already ruined. It’s no longer the same place it was in 2016 or 2010 or 2000. It’s become a corporate hellhole where you can’t get away from tracking, monitoring and other shady practices no matter what you do. Almost nothing is free and open and everything is designed to milk as much money out of the product.
AI is just accelerating this. Makes me think of the “old net” in Cyberpunk2077. Infested with rouge AI that have turned it into a battleground.
Lukewarm counter-take: That non-commercial internet is still out there.
As long as there’s nerds, there’s going to be nerds building stuff for the fun of it. Building your Lemmys, your Fediverses, your Geminis etc…
There is definitely more legislation now, dissuading some percentage of nerds, but we also have a lot more nerds…
deleted by creator
Nice! Took me to a site full of movie themes in midi format👌🏻
I got a Columbine fansite :/
deleted by creator
This just makes me feel good
Even places like lemmy or mastodon aren’t safe anymore. Everything that is public can be, and is, data mined by some corporation. There will be bot accounts or paid people pushing their agenda. And several other things.
It’s not a technical problem. We can have places that are better than the average, but the corporations will still put their tentacles in.
Even the non-commercial spaces are commercial. I’ve seen a bunch of thing disguised ads on here already. The fediverse is no different than the rest of the net. The big corps are still monitoring everything.
Every “nerd” I meet now wants out asap. The only ones who don’t seem to be pushing blockchain or AI. The tech sphere I loved is well and truly gone. The people are so profit driven nothing else can matter.
It’s just the normal capitalist cycle.
- New thing invented
- Golden age of thing
- Corporations figure out how to engineer thing for maximum profit
- Thing sucks
So its just like real life
2000 or 2010 maybe (too young to know) but I’m not sure I see any difference between now and 2016. The only major new thing I can think of is tiktok but it didn’t ruin the internet.
deleted by creator
It would be the ultimate irony if this article was, itself, generated by AI. Based on the article’s “voice”, I doubt it is, but in this brave new world…one must always wonder.
An AI writing about why it sucks? How human.
Heres another layer; AI arguing about whether an AI written article is misinformation.
TL;DR; - The internet is getting shittier at ludicrous speed, thanks to AI bots. Expect the next generation of AI to be even worse, as it’s likely to be fed the shit the current AI is dropping everywhere.
It’d be funny if a lot of bots get discontinued just because the tools can no longer discern real and fake content and just become unusable.
They already can’t. They just rely on the assumption that most of the data they collect is correct. Which is generally true, there is more correct than incorrect content on the internet. The inability of the bots to discern incorrect data coupled with their ability to make it sound authoritative is what makes them dangerous.
Who wants to bet that the powers that be will make a CAPTCHA that you have to decide if the text is written by human or AI.
Reminds me of an iconic film…
oh god, human centipede??
It’s not “ai”…it’s just a few billionaire assholes.
& it’s not just the internet. It’s the entire “reality”.
capitalism and the rich** ftfy just like they’re destroying the earth and space
The rich aren’t making spam websites full of bot-generated content. Neither is capitalism, or communism or socialism.
What’s interesting is this is also kind of a circle of tech moment too. At least for me, search has been sort of killed (or google search anyway), but it’s just back to the Internet of 1997 again, where we have “sort of useful” “search engines”, some walled gardens like AOL was, and maybe webrings or the original sort of Yahoo! curated link / subject sites / lists.
Just as bad, if not worse, than the blockchain craze that came before it.
Removed by mod
I feel like with “cheap” AI we’re entering a dark age of tech.
Eventually everything will, the tools and the tech will get more mature and sort itself out but for a couple of months/years we will be confronted with bad AI news , bad AI games, bad AI art, etc
I’m pretty confident a day will come were AI will be seen as a tool and that pure AI generated content will be seen the equivalent of a bad JavaScript game or a cheap knock clone.
deleted by creator
deleted by creator
So are all the humans.
Google made the recipy for the shitty internet cake and put it in the oven. AI is the icing.
Capitalism* ftfy
Oh no who could’ve possibly seen this coming? Literally everyone, it turns out.
Good read. I love that suburb meme as an analogy of what the internet is going to look like. IMO it’s already there.
Thank you :D I also thought it was a good analogy, especially since we’ve just accepted it as inevitable. Even with all the urbanism revival enthusiasm on the internet, they never push for beauty, just practical stuff like walkability, public transit, etc. It’s good stuff, but I want bread and roses too.
And yes, it’s getting there fast, if it’s not already there. I remember in 2015, when people still loved google, and I started talking about what I then called “The Apple Crisp Problem.” In a span of just a couple years, googling recipes went from super useful to entirely SEO blogs of maybe-not-real women in their late thirties named “Kate” taking their dog named “Pancake” to the orchard to pick the perfect apples for her also-not-real nana’s apple crisp recipe. Recipes were one of the leading indicators. Now it’s just everything. Super lame.