back
43 comments
> The social media giant said earlier this month that advertisers will be barred from using generative AI tools in its Ads Manager tool to produce ads for politics, elections, housing, employment, credit, or social issues. Ads related to health, pharmaceuticals, and financial services also are not allowed access to the generative AI features.

This seems to be a policy of preventing people from using their specific tool to generate those kinds of ads. It doesn't seem to say anything about attempting to prevent people from using other tools (say, those from OpenAI) to generate those same ads.

How would you detect that?
Detection of generated images works well, as reported by lots of works on arxiv. It's not catching them all, but lots of them. One way to understand it is that DDPMs and their underlying neural network do leave patterns in the high frequencies, often invisible to the eye. Deconvolutions typically generate low noise "marbles".
"AI detection" is a scam. Already such detection in education has been treated like "the computer says your work is AI, so that's final and you're getting sanctioned" rather than "the computer alleges you are submitting AI-generated work, can you show your work process?". It unfairly targets neurodivergent individuals who sometimes have a tendency to write in such a way that is closer to what these AI detection tools claim is AI, than neurotypical individuals.

AI detectors will (intentionally and unintentionally) end up being used in heavily ableist ways. Once it's ingrained in the zeitgeist, then we're stuck with it until it causes a scandal so big that it can't be ignored, and in the meantime it will harm millions, if not billions, of people.

AI image detection, even when it does work, will have an extremely short shelf life, as the model continue to improve. Remember when SD couldn't even reliably render hands with five fingers? Go try some image generation on the state of the art models and inspect the hands, you'll find that there are far fewer impossible hands than there used to be. This will only improve over time. What do you do when the state of the art models eventually overcome all these issues? What about when they become capable of rendering flawless text? What then?

It's like people who claim that they can tell you with 100% certainty whether an image was photoshopped. Bullshit.

False.

And if true, is only temporary.

Any "tell" can be inferred by the descriminator which will the lead the generative process towards a work around

Once again, it's not quite as simple as that.

This debate did not start just last year, lol, it has been going on for years. There's a reason you can't just simply slap a discriminator on things and get away scot free.

It'll be an ongoing cat and mouse game for....as long as we can tell, probably.

(Also, the author mentioning the deconvolution checkerboard should be a dead giveaway that this debate has been raging for years, if you've had experience in the field over the last several years.)

It's not a cat and mouse game, it a mouse game. Detection only works up to the certain point where the difference is too small to tell reliably, and the false positive rate is unacceptably high. In fact we're already at this point.
This will probably just hurt innocent people and people playing by the books. I'm a member in a small public Facebook group that gets a handful of full graphic sexual images posted daily by spam bots. If Facebook can't stop so blatant rulebreaking, how are they supposed to stop AI and deepfakes? Granted, it's easier to stop something when it's banned.
If I am reading this correctly then I think it is simply preventing the companies that buy ads on Meta from using Meta’s own generative AI features to make certain types of ads. It’s not really feasible to accurately block AI generated content anymore with good accuracy afaik.
I don't think it will hurt anyone, people will just use OpenAI or whatever other service instead of a Facebook-provided one. They are just ensuring they can muddy the waters when there is eventual blowback.
Pretty sure group type can be changed from public to private or even totally hidden - perhaps people in charge should consider tweaking visibility for a while.
True, from what I can tell, Meta is more interested in moderating different political views, than it is moderating child pornography.
I fail to see how this will “hurt” innocent people. People using ai to mislead others are not innocent. Meta should ban fake use profiles. Criminals are already leveraging procedurally generated images to milk incels of their money with fake profiles. Soon politicians and their simps will too. Ban them all.
ROFL. Yeah, ok. Meta’s intentions are totally trustworthy, sure. They get to define both “trustworthy” and “ads”.

Remember they know more about the exploitation of human emotion than any organization in human history. And remember what they did with that knowledge.

Their instinct is only to find a responsive signal. They give no consideration to consequence.

Even with the best intentions they’ll still cultivate rage because they have never discovered anything else. At this point they can never discover anything else.

Meta is an irredeemable organization. Reject them at every opportunity.

Yes, and every employee reading here should be aware that they’re complicit in enabling that horrible organization and that they never have any moral authority to engage in any argument whatsoever.
And i'm sure they'll be very successful, just as they've been successful in stopping the dozens of outright fraud announcements I absolutely don't see nearly every day, which are not at all promoting all kinds of false products, false promises and fake services through my news feed while also not at all using deeply misleading AI imagery for their ad graphics.
Good idea but how will they tell?
Meta is only preventing use of their own generative ML tools, if you use your own tools and upload the generated images they almost certainly can't tell.
They won't. They just now have another reason to ban whatever they want to without telling why.
Title looks like misinformation. Sub-title says something different (and more plausible).

Title: "Meta will enforce ban on AI-powered political ads in every nation, no exceptions".

Sub-title: "With several nations expected to hold elections next year, Meta confirms its generative AI advertising tools cannot be used for campaigns targeting specific services and issues.".

The idea of preventing advertisers from using AI at all ("no exceptions") seems fairly absurd -- for example, if an advertiser asks ChatGPT to spell-check something a human wrote, how would Meta know that ChatGPT did the spell-checking? But if Meta's just trying to manage how people access its own tools, then that'd seem like a different scenario.

Presumably they mean that their tools couldn't be used directly, rather than not at all, though. For example, if Alice uses one of Meta's tools for some other declared purpose to spell-check a word that Alice'll then include in an ad that she'll ask Bob to deploy on Meta, then how would they detect such indirect usage? Though they might still try to detect their own tools' signatures on, say, images or longer bodies of text.

Coming soon: AI detector detectors.
These kinds of arms races always remind me a particular scene from The Big Hit: https://youtu.be/Iw3G80bplTg
Sure, AI detectors all the way up.
1947 repeats itself.
They are yet to define their meaning of the word "exception".
> Meta will enforce ban on AI-powered political ads in every nation, no exceptions

Is today the 1 of April ? And they will let selective political ads further, isn't it ?

This seems like a very strange thing to enforce. Why ban a tool for creating ads as opposed to problematic kinds of ad content?

For example, if I want to make a generically patriotic ad, and I generate a picture of an eagle holding a flag to use in an ad, that doesn't seem like an issue. On the other hand, if I want to make an image of Biden walking with a cane and do so through Photoshop, that does seem like an issue.

This maybe stops candidates for low offices with small campaign budgets from using cheap/free AI tools to do fakery, but anybody running for major office has more than enough money to hire someone to make whatever they want with Photoshop.

Glad to see that, with no corporatist boycott or so-called whistleblowing from groups of "beach friends," Meta is taking appropriate steps against the disinfo -- i.e. propaganda -- that is actually real and actually causes harm. Let's just hope this time around they don't let corrupt FBI agents unilaterally delete links to news articles either.
I'd be cooler if they banned all political ads.
I’m taking bets that there will be exceptions.

Seriously, any takers?

Images? Text? I have no idea how they're going to reliably ban AI-generated or AI-inspired text and I really don't think images are the limiting factor for political ads.
I suspect enforcement will be arbitrary.
"X/Twitter will discount AI-powered political ads in every nation, no exceptions."
(But just the right wing stuff)
For other brothers who are by character and upbringing on the right side, but want to and put work in understanding and keeping connection with the left side, I'd recommend this Jordan Peterson's podcast episode: 395. Difficult Conversation as the Precondition to Progress | Adam Smith. It helped me understand the steelman version of argument behind diversity, equality and inclusion, which are often ridiculed in the traditional right wing bubble.
There really can be no meeting in the middle to the left because the ideas being opposed by the right are those that involve fundamental rights for those groups.
The "fun" thing about this sentence is about the same number of people will vehemently agree with it as it is and if you swap "left" and "right".
I don’t think it’s swappable. No one on the right is losing fundamental rights due to the lefts views. They might see a decrease in the quality of their lives (reduced privilege, sharing job options with immigrants etc.), but it’s not a fundamental attack to their personhood.
the fun thing about that argument is it can be applied to creationism vs. science, too, but we know that one of those is wrong and the other right

turns out "a minority of people with me" isn't that convincing of an argument

"fundamental lefts"?
And that's perfectly fine. Not everything needs to be a "meeting in the middle" or a "compromise". Ideas need not be watered down.
Except for the ones that they themselves employ :)
I will enforce a personal ban on use of all Facebook crap, no exceptions.