Slashdot is powered by your submissions, so send in your scoop

 


Forgot your password?
Close
typodupeerror
×
Google AI Advertising

Did Google Fake Gemini AI's Output For Its Super Bowl Ad? (theverge.com) 40

Posted by EditorDavid from the cheesy-ads dept.
Google's Super Bowl ad about a Gouda cheese seller appears to be using fake AI output, writes the Verge: The text portrayed as generated by AI has been available on the business's website since at least August 2020, as shown on this archived webpage. Google didn't launch Gemini until 2023, meaning Gemini couldn't have generated the website description as depicted in the ad.
The site Futurism calls the situation "beyond bizarre," asking why Google doesn't seem to trust its own technology. Either Google faked the ad entirely, or prompted its AI to generate the web page's existing copy word-for-word, or the AI was prompted to come up with original copy and instead copied the old version. In the publishing industry, that's referred to as "plagiarism."
And ironically if Gemini did plagiarize that text, the text that it plagiarized is also inaccurate.

Did Google Fake Gemini AI's Output For Its Super Bowl Ad? More | Reply

Did Google Fake Gemini AI's Output For Its Super Bowl Ad?

Comments Filter:
  • Who really cares, because it's a commercial and is designed to manipulate people into believing the product is better than it really is. Assuming that it was a different example do you think it wouldn't be one where the product appears incredibly capable as opposed to an example where it fails miserably? Further, is it at all surprising that an LLM might actually wholesale plagiarize something instead of remixing it? Ask a specific enough question and you can almost guarantee this as I discovered last year

  • Of Course It's Faked (Score:5, Informative)

    by Deep Esophagus ( 686515 ) on Sunday February 09, 2025 @10:52AM (#65153581)

    Everyone in IT knows that AI is just a Clever Hans pony trick, an illusion to fool the rubes who don't know better. It doesn't think, it doesn't reason; it just digests and regurgitates without any conscious understanding of what it's doing. At my company we are encouraged to use Copilot for increasing productivity, but it slows me down. I ask it for API endpoints, and it just makes crap up. I ask it for Powershell code, and it invents nonexistent functionality. Whether it's AI-generated art, or prose, or code, or ad copy, or music - you're going to get garbage you have to manually correct.

    • Re: (Score:2)

      by MpVpRb ( 1423381 )

      Agreed, for consumer facing AI
      Alpha Fold and other projects using AI tools in research are real

    • You're not wrong, but you should know that Copilot is behind the foundational models.

      And using Copilot as a benchmark for anything isn't an accurate representation of what's possible right now.

    • This is false: "Everyone in IT knows that AI is just a Clever Hans pony trick". I mostly know computer scientists and senior software engineers, but I would wager not even a majority of people in IT have such a stupid opinion.

      • Yeah it's frustrating to see that being scored as Informative. AI generated code isn't perfect but it's clearly faster/better than googling the forums and tutorials.

        • I think that depends a lot on how effective your queries are, how quickly you can skim the results, and how quickly you can take separate concepts and put them together into one cohesive one. It just seems to me that a lot of people are really bad at all three of those and those are the ones praising AI.

          • I think that depends a lot on how effective your queries are, how quickly you can skim the results, and how quickly you can take separate concepts and put them together into one cohesive one. It just seems to me that a lot of people are really bad at all three of those and those are the ones praising AI.

            It's like saying people that are bad at joinery are the ones praising screws.

            I know what you're trying to imply, but the guys complaining about screws splitting their wood aren't good joiners and good joiners use screws exceptionally well.

    • Everyone in IT knows that AI is just a Clever Hans pony trick, an illusion to fool the rubes who don't know better. It doesn't think, it doesn't reason; it just digests and regurgitates without any conscious understanding of what it's doing. At my company we are encouraged to use Copilot for increasing productivity, but it slows me down. I ask it for API endpoints, and it just makes crap up. I ask it for Powershell code, and it invents nonexistent functionality. Whether it's AI-generated art, or prose, or code, or ad copy, or music - you're going to get garbage you have to manually correct.

      It's just another tool.

      If you can't use it usefully, that's on you. You are either using the wrong one, or using it wrong/ineffectively.

      Lots of us use LLMs every day in programming and it's a very useful tool.

      (It's like listening to someone who whittles all their fasteners and joints tell me that hammers don't work. I mean yeah, everything isn't a nail, but ... you're still kind of whacked.)

    • Re: (Score:2)

      by mjwx ( 966435 )

      Everyone in IT knows that AI is just a Clever Hans pony trick, an illusion to fool the rubes who don't know better. It doesn't think, it doesn't reason; it just digests and regurgitates without any conscious understanding of what it's doing. At my company we are encouraged to use Copilot for increasing productivity, but it slows me down. I ask it for API endpoints, and it just makes crap up. I ask it for Powershell code, and it invents nonexistent functionality. Whether it's AI-generated art, or prose, or code, or ad copy, or music - you're going to get garbage you have to manually correct.

      Forget AI, everyone in advertising knows it's completely fake and most of us who aren't in advertising also know it. That's why the burger you get at Wendys/McDonalds/et al. looks nothing like the picture on thead on the bus shelter outside. They used to have means of faking it (I.E. glue, paint, et al) but now just use CG.

      Any advertisement indistinguishable from magic is insufficiently faked.

    • Re: (Score:2)

      by Fross ( 83754 )

      I think that may be a you issue.

      I've been coding for 40 years, and professionally for 30, and using tools like this speeds me up a lot, I'd say 25-30% more good code generated at the end of the day. Sometimes to generate code, sometimes to rubberduck or sift through sources for me. Certainly reasoning models are getting better at it. I trust it for low/medium complexity tasks - not blindly. But easily as much as I would a junior.

      It doesn't get it 100% right the first time, every time. But if you can't debug

    • My power drill keeps camming out and doesn't know where the holes are supposed to go

      Everyone in IT fits in two categories, people that can figure out productivity enhancing tools, and those that don't.
      That sounds mean, look, I know we're all up to our eyeballs with tools of questionable quality that promise some wonderful bullshit, and many have completely unreasonable demands for cognitive focus. I get that.

      LLM is one of the good ones, like learning regex for example. Yah, they uh.. "cam out" so to speak. You need to know how to use them. Yes, that does sound exactly like every excuse for

  • Mankind's quest for the ultimate slave is having issues it seems.

  • Who would have thought the AI that depicted Vikings and Nazis and being Black would have faked output. I'm shocked! Shocked to find that fake output is going on in here.

    • Who would have thought the AI that depicted Vikings and Nazis and being Black would have faked output. I'm shocked! Shocked to find that fake output is going on in here.

      I tried one of these new video generating AI tools to see what it could do. I created a short 1 minute video about German Shepards and at the ned it talked about walking your Shepard while showing someone walking a lab. Of course, that's not stopping people from hyping AI video as the new way to make easy money on YouTube.

  • Why ... (Score:3)

    by PPH ( 736903 ) on Sunday February 09, 2025 @11:16AM (#65153625)

    ... Google doesn't seem to trust its own technology.

    Do you?

  • As the cereal box says... (Score:3)

    by fuzzyfuzzyfungus ( 1223518 ) on Sunday February 09, 2025 @11:43AM (#65153679) Journal
    They didn't 'fake' the output; it's just a 'serving suggestion' that has been 'enlarged to show texture'. All on the up and up.
  • What are you doing? Next they will think it isn't sentient : P

    • What are you doing? Next they will think it isn't sentient : P

      Remember, it has to be centient before it can be dollarient.

  • I thought the guy in the toothpaste ad is a dentist!

  • In the LLM space, nobody can do except for very simple stuff with low quality requirements.

Slashdot Top Deals

How come financial advisors never seem to be as wealthy as they claim they'll make you?

Close