Hacker reverse engineers his Samsung TV and finds on-device AI listening for your dog and baby, watching what you watch, and borrowing your soundbar's mic if the TV doesn't have one. Then OpenAI banned him for reverse-engineering it.

Source:

you are viewing a single comment's thread
view the rest of the comments
[–] 29 points 5 days ago (59 children)

Once again, theres an amazing amount of ethical ai use out there but apparently nigh nobody cares about that

  • source
  • parent
  • hideshow 59 child comments
  • [–] 42 points 5 days ago (37 children)

    Ethical from the start you mean? Trained on data with the blessing of the owner?

  • source
  • parent
  • hideshow 37 child comments
  • [–] 16 points 4 days ago (4 children)

    First of all, yes, and also not all ai has to be generative or llms. Theres also a lot of other ai that was done before openai fucked everything up.

  • source
  • parent
  • hideshow 4 child comments
  • [–] 16 points 4 days ago (2 children)

    When we say AI, we generally mean the LLMs. It's become a marketing term for them.

    Obviously, we used neural networks and such ethically for a long time.

    The problem isnt those, it's the mass harvesting of data, the endless copyright breaches, the stealing of peoples work. All to resell it. While ruining the planet environmentally, economically and it has fucked with the way people think.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 9 points 4 days ago (1 child)

    I mean yeah thats true but ethical llms are still completely possible but nobodys investing into it

  • source
  • parent
  • hideshow 1 child comment
  • [–] 12 points 4 days ago

    Because it's about money. That's it.

    There's no money to be made from ethical ones, you need enough people willing to give you the rights to their stuff for free, because you would not be able to give royalties to them all.

    We already see how unsustainable it is currently with the grand theft models and are just waiting for the bubble to burst.

    AI can fuck off and die, we dont need it.

  • source
  • parent
  • [–] 8 points 5 days ago (21 children)
  • [–] 10 points 4 days ago (20 children)

    Well, their books are, yes.

    But, the similarities largely stop there, in this instance. Training "AI" on literature isn't at all similar to a single person checking out a book/magazine/dvd/et al, and that's not even getting into the fact that the aforementioned fuckers are destroying tons of physical copies IRL to "train" their bullshit bots.

    "AI" ≈ libraries? Pff. We would be so lucky. 🥲

  • source
  • parent
  • hideshow 20 child comments
  • [–] 9 points 4 days ago (19 children)

    When y'all complained about them leeching Anna's Archive, the legal approach you demanded means buying and shredding one copy of any book they use. They can't resell the book or share their scan because that's piracy, which Lemmy cares deeply about in this context and only this context.

    Training is transformative. A program that can retell Shrek in the style of The Hobbit is not even the same category of thing as either original work. What a network is and does cannot be a substitute, unless the Silmarillion had more swamp ogres than I've heard.

  • source
  • parent
  • hideshow 19 child comments
  • load more comments (19 replies)
  • load more comments (7 replies)
  • [–] 12 points 4 days ago (7 children)

    The 10% of ethical AI use is drowned out by the other 90% of absolute liquid trash.

    That's not touching the ethics of stealing everyone's work for your AI then selling it back to them.

    That's not touching the ethics of aggressive Data Center build outs that threaten people down to their actual health.

    That's not touching the fascist supporting corporations that run and own the frontier AI models.

    That's not touching the lack of those frontier AI company's ethics. They operate like it's the fuckng wild west.

    How can you possibly have a victim complex about this?

  • source
  • parent
  • hideshow 7 child comments
  • [–] 7 points 4 days ago (1 child)

    That’s not touching the ethics of stealing everyone’s work for your AI then selling it back to them.

    Yeah, it's the "selling it back to them" part that gets me. I don't want to outlaw scraping, because then the Wayback Machine couldn't function. The Wayback Machine is free, though. It's a public service. (One which is very much needed, IMO.) If, instead, the Internet Archive required that you pay for a subscription in order to use it, I think most people would be very upset, and rightfully so.

    You want to scrape the Web to feed your LLM? Cool! Make it an open-weight model.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 5 points 4 days ago (1 child)

    That's also not counting the ethics of barraging every site on the internet with a DDOS-level surge of webcrawlers scraping up data for AI training, to the point where every significant website out there has to implement anti-AI countermeasures in order to just stay operational.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 7 points 4 days ago

    There's so much bullshit with LLMs that needs to be cleaned up, regulated, and outright outlawed before I'm gonna sit here and praise these chucklefucks for opening a kind of Pandoras box with basically no thought.

  • source
  • parent
  • [–] 1 point 4 days ago (2 children)

    I wonder what % of actions being ethical would be needed from Germany before WW2 would be considered unjustified.

    I'm willing to bet that 10% of the actions the Nazis did (note: out of all the actions the German govt did, including enforcing laws, running trains/buses, collecting taxes, building infrastructure etc) were ethical, and yet the Nazis were still obviously the bad guys in WW2.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 3 points 4 days ago (7 children)

    This comment doesn't even make sense in the context? The danger lies in unethical use. It's also unregulated.

    Don't forget, there was a very brief time period when Google was ethical. How long did that last?

  • source
  • parent
  • hideshow 7 child comments
  • [–] 5 points 4 days ago (5 children)

    I think the point is people like the person hacking Samsung smart TVs with AI should get some credit, purist "AI bad" takes don't account for this. Also if there's regulation it's likely to start by targeting good uses like that one. Fuck OpenAI though.

  • source
  • parent
  • hideshow 5 child comments
  • [–] 2 points 4 days ago (1 child)

    1 good thing coming of it, doesn't wipe away all the fucking harm AI is causing. If AI never blew up, the world would be better off. It's introduced a whole host of problems we didnt need, just for the sake of a few people getting richer.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 1 point 3 days ago*

    1 good thing coming of it, doesn’t wipe away all the fucking harm AI is causing. If AI never blew up, the world would be better off. It’s introduced a whole host of problems we didnt need

    So what though? You don't have the option of getting rid of it, and it's a technology, not a person. Nuance is called for, especially if the alternative is refusing to acknowledge good things people do with it, or focusing condemnation on "AI" as a monolithic entity instead of the particular people and companies doing harm with it. Case in point, OpenAI banning this person.

  • source
  • parent
  • [–] 1 point 4 days ago (2 children)

    I'm sorry, but AI generally remains "bad" until regulation takes place, community concerns about data centres are properly considered and protections are put in place. Right now, billionaires and corporations are running amok and that usually doesn't end well.

  • source
  • parent
  • hideshow 2 child comments
  • load more comments (3 replies)