Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
AI Facebook

Meta's New Rule: If Your Political Ad Uses AI Trickery, You Must Confess (techxplore.com) 110

Press2ToContinue writes: Starting next year, Meta will play the role of a strict schoolteacher for political ads, making them fess up if they've used AI to tweak images or sounds. This new 'honesty policy' will kick in worldwide on Facebook and Instagram, aiming to prevent voters from being duped by digitally doctored candidates or made-up events. Meanwhile, Microsoft is jumping on the integrity bandwagon, rolling out anti-tampering tech and a support squad to shield elections from AI mischief.
This discussion has been archived. No new comments can be posted.

Meta's New Rule: If Your Political Ad Uses AI Trickery, You Must Confess

Comments Filter:
  • Yeah, sure. (Score:3, Insightful)

    by Black Parrot ( 19622 ) on Monday November 13, 2023 @10:28AM (#64002239)

    Like all social media, they'll interpret the rules to give the owner what he wants.

  • Democracy is over (Score:2, Interesting)

    by DarkOx ( 621550 )

    So if you have the backing of big corporation like MSFT or META, or Soros, or Peter Thiel to spend on high production value advertising, you can make whatever you want, say whatever you want.

    If you don't have money to build sets and pay actors but try to do an ad on the cheap with AI - well you'll have all your posts labeled with some kind of mark of taint, which will be universally seen as 'you're a liar' no matter what the content actually is.

    • Relax it's fine (Score:5, Interesting)

      by rsilvergun ( 571051 ) on Monday November 13, 2023 @10:50AM (#64002293)
      California is starting to teach classes in recognizing fake news. Aka "Critical thinking and how to evaluate claims". Every blue state in the Union will do it soon after The red states will drag their feet because, no joke, religious objections [austinchronicle.com], but pretty quickly people will adapt.

      Money in politics isn't about lying, it's about name recognition. People vote for the name the recognize, at least the "independent" voters do. The rest tend to vote on party lines first and finally issues come into play. That isn't great, but it means fake news isn't going to have as big an impact on Democracy as folks think.

      Meanwhile a better educated population means that people will do a better job evaluating candidates, so you'll see less of that name recognition voting going on. The parties are so starkly divided you'll still have party line voting, I don't think anyone who'd consider voting for a Democrat could imagine crossing party lines at this point. The talk of closing borders at states to prevent abortions alone shuts that down and purity tests in the Republican party around abortion mean you can't win a primary election without supporting that. But that's still at least issues based voting.

      As the baby boomers age out of voting and a younger generation with a *lot* less property and money comes of age you'll eventually see a new new deal though and things will get better. The only question is if something like "Project 2025" (google it) will install the next GOP president as dictator. That's a real possibility, but it won't be fake news causing that, it'll be a classic soft coup scenario.
      • Trust in media sources - and eyeballs - has/have precipitously declined and will continue to do so. There was a recent Pew poll on this very topic. The trust levels are in the sub-20% at this point in several demographics.

        As Goebbels conveniently told us, there are limitations on the power of propaganda. He found that even with a healthy dollop of the truth mixed in with the propaganda, it was disbelieved anyway.

      • by dfghjk ( 711126 )

        "That isn't great, but it means fake news isn't going to have as big an impact on Democracy as folks think."

        Right, because "fake news" can only impact Democracy ON ELECTION DAY, definitely not before. Couldn't possibly impact who turns out to vote, much less how they vote as you say. Some real deep thinking here.

        You see, politicians gave up on telling lies because it had no impact on elections.

      • Blue state politicians pump out more fake news and propaganda than all others.

      • Youâ(TM)re delirious if you think that the average American is able to (or better - willing to!) distinguish fake news from real news.

        Education in the US is a jokeâ¦
      • If you are citing the Austin Chronicle as a source of truth, that says a lot lot more about you than your supposed enemies. What a dolt.
    • It will be as useful as “serving suggestion” or “your mileage may vary” disclaimers.
    • Re:Democracy is over (Score:5, Informative)

      by ArchieBunker ( 132337 ) on Monday November 13, 2023 @11:11AM (#64002343)

      You said it, democracy is over.

      https://apnews.com/article/ohi... [apnews.com]
      https://www.cincinnati.com/sto... [cincinnati.com]
      https://news.yahoo.com/having-... [yahoo.com]

      • Those idiots in Ohio trying to block something to pass with a 13% margin of victory are going to get cut to pieces in 2024. Even gerrymandering won't save them because gerrymandering lets you win by 2 to 3%, 13% not so much.

        I can tell you right now every Republican strategist is desperately telling them to just accept the defeat and to campaign on it. That's the reason abortion exists is an issue it was specifically selected by the Republican Party in the 60s as a wedge issue that could be used to divid
    • by dfghjk ( 711126 )

      You think it will be that good, huh?

    • My question is, why would anyone outside a political campaign be generating political ad content for that campaign?

      • My question is, why would anyone outside a political campaign be generating political ad content for that campaign?

        Why not? Haven't you heard when flyers are sent to voters on the weekend, or even the day, before an election which have no indication who they came from, but will either back one candidate or the other or more usually, denigrate one candidate including lying about them.

        Whether the flyer came from the candidate's campaign is irrelevant. The point was to make it so voters got one last thing to think about before the election, but too late for the campaign to respond.

        To your questions, imagine someone uses A

  • by backslashdot ( 95548 ) on Monday November 13, 2023 @10:33AM (#64002245)

    I didn't chop the cherry tree. It was Thomas. I cannot tell a lie.

  • People were faking videos before AI technology was a thing, using old-fashioned methods. Look at the famous video of Obama kicking the door. It's not really him, but was created using a lookalike actor, similar stage, clothing, and props, and a simple crossfade where the two clips meet.

    • That's just it though, that little clip took some production and care and thought. Now I can just go into Midjourney, "Obama walking up to a door and kicking it down" and it will spit me out something that probably is not as high quality but close enough to fool people browsing on Facebook which is a real problem even for something innocuous like that.

    • We've had Photoshop and doctored images for decades too.

      The problem is the same that it's always been: media literacy, or lack thereof. At the end of the day, it serves each person to remember that images, video and audio are but one data point and that they can be misleading because they don't paint the full picture.

      Even without the use of AI or lookalike actors, videos can be edited using un-doctored source clips in such a way as to mislead. By leaving out crucial context or by editing to create a differe

  • Having to disclose if AI was used to touch up a video or the vocals seems like overreach. If the use of AI is being used to mislead, that is one thing. But if they are using AI to remove a facial blemish or change the color of the candidate's tie, I don't think a disclosure if necessary.

  • by Petersko ( 564140 ) on Monday November 13, 2023 @10:43AM (#64002265)

    "You will constrain your modes of dishonesty to well-established patterns of evasion, deception, and outright bald-faced lying. Don't get clever. That's Meta's job."

  • If the findings are not in agreement with the reader's "Trusted Source of Truth" then they will be dismissed as "Fake Newz", "MinTruth", or whatever else will allow them to maintain their particular world view.

    If the facts don't match up with their viewpoints, they'll insist that the facts are the problem.

  • by rsilvergun ( 571051 ) on Monday November 13, 2023 @10:59AM (#64002313)
    just put a little easily ignored notice on them? Great. Good job.

    I don't think this stuff effects elections as much as people think. Name recognition still Trumps (pun?) fake news all day long.

    But this kind of crap can set off a random lunatic on the right wing to go off shooting people. That's happened more than once. Heck it seems to happen 3 or 4 times a year, with another dozen or so cases where they pick the guy up before he manages to do the shooting.

    That said, I saw a little graphic today with a list of Trump quotes and Hitler quotes. It wasn't "can you tell the difference" like the old memes, it was a list of things Trump said that were 1 to 1 or nearly 1 to 1 out of old Hitler speeches. Meanwhile we've got things like "Project 2025", which is a plan to disable the bureaucracy that got in the way of Trump overturning the last election (among other things, like seizing control of state legislatures so that fake electors can be sent to vote in whatever president without pesky voters getting in the way). Oh, and there was that thing Ron DeSantis said about slitting government employee's throats ("metaphorically" of course)...

    These are pretty major things that are getting little or not mainstream media coverage. At this point it's pretty clear the media is ignoring these things, and you have to start asking why.

    And then you look at how much ad revenue they're anticipating next year. It's around $2 billion dollars all told. And it all makes sense. It's gotta be a horse race, otherwise who's gonna put up all that cash for advertisements.

    Basically, our entire democracy is at risk because of ad revenue. Your ad blocker might save the world! Huzah!
    • by dfghjk ( 711126 )

      You're going to talk about Hitler similarities with Trump right after dismissing the power of "fake news"? You realize that Hitler coined the term "fake news" and "the media is the enemy of the people", right? Propaganda was Hitler's most powerful weapon in his rise to power, and it was Trump's as well.

      "Basically, our entire democracy is at risk because of ad revenue."
      Well, no. Too reductionist. And Elon Musk's purchase of Twitter is about future propaganda directly benefiting the owners, no ad revenue

      • Hitler didn't coin the term "fake news".

        Abundant famous examples of deceitful messaging that existed in antiquity can be found by even the laziest researcher using nothing more than Wikipedia, which might be widely seen as a poor source, but the citations on the articles are real. But sure, nothing between antiquity and Nazi Germany had a name for that, and when it was named, it was named by Hitler, and when he named it he used the English language.

        Be serious instead of a timewaster. Had you said Goebbels I

        • Hitler is credited with coining the term "big lie" however (wikipedia):

          A big lie (German: grosse Luge) is a gross distortion or misrepresentation of the truth primarily used as a political propaganda technique.[1][2] The German expression was first used by Adolf Hitler in his book Mein Kampf (1925) to describe how people could be induced to believe so colossal a lie because they would not believe that someone "could have the impudence to distort the truth so infamously". Hitler claimed that the technique ha

          • Big Lie is what Dick Chaney & Bush Jr did. It's telling a lie so crazy that nobody believes you would tell it. "Iraq has WMDs" is the classic modern example.

            Repetition like you're talking about isn't Fake News either. It's repeating a lie until it sinks in. That's classic marketing. It's called an "Ad Blitz" these days.

            "Fake News" is a slogan used to say there is no truth but your own. It's what religious extremists call "God Glasses". It's new, at least in politics. It's like double think on st
        • More to the point, Trump didn't re-invent "fake news" or borrow it from Hitler, he co-opted it from the modern media.

          A big issue during the 2016 campaign was literal fake news sites. These sites were largely right wing sites fabricating stories out of this air, and these stories would get circulated around social media.

          When the discussion of this started to hit the mainstream media Trump recognized the catchiness of the phrase (and the risk if it got associated with his supporters) so he simply started usin

    • by e3m4n ( 947977 )
      Yes because no left-wing nutjob ever started shooting people. Certainly not ames Hodgkinson, Aiden Hale (tranny name), not James Holmes, etc. Your wife must hate you in an argument being all 2-faced. Maybe DC comics will look you up for their next Batman villain.
  • Who are these people?
    Who asked them to decide who can say what?
    How the hell is this legal?
    Why aren't our elected officials doing something to rein them in and tell them to stay in their place? (Hint: that last one is a rhetoretical question...)

    Facebook and Microsoft are the gatekeepers of political integrity on the internet now. The very same sumbitches who inflict AI upon the world and stop at nothing to steal our private data. The mind boggles.

    God this dystopia is depressing...

    • You are using their product, if you don't want to be policed don't use Facebook. You can't walk into a church and start preaching from the Koran.

    • by dfghjk ( 711126 )

      "Who are these people?"
      People who fear being held accountable for the damage they do for profit.

      "Who asked them to decide who can say what?"
      No one. They do it in response to threat of accountability.

      "How the hell is this legal?"
      Says a LOT about you. How is it legal that they do what they want with the private platforms they control?

      "Why aren't our elected officials doing something to rein them in and tell them to stay in their place? "
      They are, they're the ones threatening accountability.

      "Facebook and Micr

    • by HiThere ( 15173 )

      It's legal because it's their playground, so they get to set the rules. This isn't a government regulation.

      Now whether it *should* be regulated by government, that's a different question. But corporations are allowed to control what text/graphics appears on their web pages. (So are people, but few people have popular web pages of their own.)

      • It's legal because it's their playground, so they get to set the rules. This isn't a government regulation.

        It's funny how people are quick to acknowledge and accept modern fiefdoms.

        I would agree with you if there was an alternative place for public speech to take place. But the reality is, those privately-owned and privately-run platforms now *are* the de-factor space for public speech. That includes the press, radio and television, which are also owned by a frighteningly small number of ultra-wealthy individuals.

        As such, in my opinion, they lose the right to dictate what goes on on their platforms: they wanted

        • by HiThere ( 15173 )

          You're talking about how the laws SHOULD be written, no how they were written.

          OTOH, do you want the government deciding even more than they do what can be told? Writing the laws for how things should be done is a very difficult process, and I don't see anyone with any power that I'd trust to alter the basics.

        • by dryeo ( 100693 )

          Well, it's an improvement from when, if the powers that be didn't like what you were saying in the public square, they would have you arrested (loitering, disturbing the peace, vagrancy etc, followed by resisting arrest), held in jail for a couple of days, beaten and run out of town.

  • Also needed (Score:5, Interesting)

    by necro81 ( 917438 ) on Monday November 13, 2023 @12:06PM (#64002463) Journal
    Leaving AI aside for a moment: one thing that I would like all political advertisers to be required to do is have one central location where every advert they produce gets posted for all to see. With Meta and Google able to slice and dice every demographic to a fair-thee-well, the ads that one group sees are likely to be quite different to other groups. Even mailings can be very carefully tailored to a particular audience these days. This is different to bygone years of mass media, where all you had to do was turn on the TV to see everyone's ads. Everyone seeing everything made things a bit more above-board, and less (though not complete) susceptible to lies and fabrication.

    The rule would apply to any organization engaged in political advertising: campaigns, obviously, but also PACs, political parties, 501(c)3s and 4s, unions, corporations, etc. (Individuals would be exempt.) Each organization would be required to have a single clearinghouse where one could find the original materials: email templates, text messages, videos, glossy mailers, etc. Ideally, the details of each ad campaign would be included as well, e.g., this was a Facebook advertisement, targeting categories A, B, and C, from June 29 thru July 17. Internal messaging would also be exempt (e.g., an email from a union to union members; a text from the RNC to registered Republicans), in part because those tend to get seen by the public sooner or later anyway.

    Furthermore, make the outlets that broadcast such messages (TV, radio, Facebook, Google, ad clearing houses, mass mailers) liable ($ per ad instance) if they themselves did not verify that the political advertiser had posted the original content to their "everyone can see" clearinghouse.
  • by Chaseshaw ( 1486811 ) on Monday November 13, 2023 @12:08PM (#64002475)
    The implication being political ads have been honest and forthright up until now? Is everyone at Facebook too young to remember the Daisy Ad in 1964? (or, for that matter, whatever their education, it didn't include a political science class?) The answer to this doesn't have anything to do with AI, or attack ads, honesty or politics. It's to learn for yourself and to teach those around you NOT to be swayed by sensationalist fear-mongering. Learn your parties. Learn the issues. If you're casting your vote out of fear for what "the other guys" will do, you're part of the problem. Democracy is "the rule of the people." Fixing it starts with YOU. Not "them", not a company, not a policy.
    • The implication being political ads have been honest and forthright up until now? Is everyone at Facebook too young to remember the Daisy Ad in 1964?

      Considering they would have to be born in the early 50s, I'd say the answer to your question has already been answered. I mean, duh.

      Forget implications. The money-based motive here is simple; abuse AI as a bullshit excuse to dismiss and otherwise ignore the obvious human-based corruption that has happened in politics since long before Daisy was running ads, while continuing to profit heavily from such advertising in an election year.

      The "other guys" argument is merely another Weapon of Mass Distraction.

    • by GlennC ( 96879 )

      Learn your parties.

      There's only one, so that's easy.
      It has the "bickering spouses" of Team Red and Team Blue.

      If you're casting your vote out of fear for what "the other guys" will do, you're part of the problem.

      Unfortunately, the Party has adopted this as one of their core values.
      They WANT people to keep the focus on Team Blue versus Team Red because they win either way.

      Fixing it starts with YOU.

      This assumes that democracy in the United States can be fixed.
      I see no evidence of this, although I welcome any that others may have.
      As always, I refer you to my signature.

  • "AI" is a misnomer getting popularized because of models where you ask for things via sentences rather than buttons however it's still just an algorithm. There isn't a clear definition of what "AI" is defined as. Even an image filter changing an image to black and white could be considered "AI" under the standards issued by Biden's new Executive Order on "standards for AI safety and security, " which I'd guess is what is triggering Meta to take this action. These types of algorithms have been in use for dec
    • by HiThere ( 15173 )

      Sorry, but MidJourney and other LLMs are AI. They aren't AGI, but they're AI. So is/was Sargon. It was basically a tree parser with alpha-beta pruning, but it was/is AI. (Not a very advanced one, but what do you expect of something running on the Apple ][ in the early 1970s.) The first AI program was probably Samuel's Checkers player. https://medium.com/ibm-data-ai... [medium.com] Back in the early 1960's Scientific American had an article (Martin Gardner?) about an AI computer built with matchboxes, string, jujub

      • See that's where you and I differ. I don't consider literally any response to an input AI. AI i assume has to have some sentience and at minimum not just respond to prompts. Classifying most of these things as AI would be like saying Microsoft Word is AI because it takes a key press and turns it into a letter on the screen. Even having 'parameters' created by training data doesn't make it AI. AI is just used because it's more interesting than saying algorithm.
        • by HiThere ( 15173 )

          If it can learn and adapt to the input it encounters, it's an AI. Existing AIs aren't very smart, but that's just a matter of degree. And there's no theoretical reason to put people at the top of the possible AIs, besides that they don't qualify on the "artificial" bit.

          • And word has spell check. That doesn't make it AI. AI is completely a promotional term for 99% of these products.
            • by HiThere ( 15173 )

              Does the "word spell check" learn the way words are spelled without manual intervention? If so, then it's a rudimentary AI, but I really doubt that it does.

              • Yes, it learned it from a database of words, like all modern AI's work and it responds to "prompts" of letters and make suggestions of how you should finish your word or change your grammar. I think you are assuming that AI in these examples is continuously learning which is not the case. They "learn" aka are programmed with specific training databases.But whether you or the next person continue to ask it questions it doesn't care. Sure maybe it will take into account your history of questions or your spec
  • Meta's New Rule: If Your Political Ad Uses AI Trickery, You Must Confess

    Common F. Sense's New Rule: If you must abuse "AI" as an excuse to dismiss obvious human trickery in politics, you must confess.

    Meta likes pretending us "dumb fucks" don't realize how much AI has fuck-all to do with corruption in politics, now or before.

  • Since Photoshop now uses AI...

  • by Kelxin ( 3417093 ) on Monday November 13, 2023 @02:57PM (#64002987)
    USA passes laws that all illegal gun owners must confess that they have illegal guns. Yeah uh-huh. Everyone is going to do that, right?
  • So you can still use it, just have to confess? ...like...with a microprint disclaimer?.... riiiiight... self policing has always worked historically

Think of it! With VLSI we can pack 100 ENIACs in 1 sq. cm.!

Working...