Re:Introducing Pangram Image 1.0
by my002 on Thursday November 05, 2026 @04:47PM
(#50384922)
People often shit on AI detectors but Pangram's text AI detection has been very good, IME. Not perfect, but good enough to be useful in most circumstances.
Re:Introducing Pangram Image 1.0
by jszymborski on Thursday November 05, 2026 @07:20PM
(#50386505)
My trouble with Pangram is that the calibration is totally off. It can tell you if there is a drop or an ocean of AI in the text you submit. The "X% of AI" message doesn't really mean much. You can test this yourself: Slowly add generated text to a human write paragraph and you'll see it hit "100% AI" quite quickly.
Re:Introducing Pangram Image 1.0
by Matticus_Rex on Saturday November 07, 2026 @01:54AM
(#50402277)
That's the state of the tech right now. It'll be nice if we can get it and keep it more granular than that, but even at the current state it's pretty impressive.
Re:Introducing Pangram Image 1.0
by xnx on Thursday November 05, 2026 @05:43PM
(#50385570)
I'm very skeptical of AI detection and watermarking for text (especially short passages), but AI detection for images and videos seems much more plausible (for now at least).
I would love if Meta licensed this tech to use for everything on Facebook, Instagram, and Whatsapp. Including ads!
Re:Introducing Pangram Image 1.0
by dwohnitmok on Thursday November 05, 2026 @08:03PM
(#50386852)
I don't know about Pangram's viability as a discriminator but I can tell you as someone who has recreationally trained GANs that they're really really finicky to train and training fails in all sorts of inscrutable ways. It wouldn't surprise me if even though in theory you could do a GAN-style form of training with Pangram, in practice it might actually be very difficult.
Re:Introducing Pangram Image 1.0
by treyd on Thursday November 05, 2026 @10:55PM
(#50387955)
Modern diffusion models aren't trained like GANs. GANs failed to scale and generalize. Diffusion models do sometimes use Variational Auto-encoders (VAEs), which are a similar structure but which functions differently.
Re:Introducing Pangram Image 1.0
by program_whiz on Friday November 06, 2026 @06:27AM
(#50391040)
Also you aren't getting solid gradients, it would be a simple yes/no (or confidence score), as you don't have their model. GANs failed to generalize, even with full access to the discriminator model (providing robust gradients to all weights), its extremely difficult to train in what is basically a reinforcement learning loop over a possibly extremely spiky, nonlinear terrain.
Think of the resources it takes to train good text generators, with perfect knowledge of the next token to generate. Now imagine to trying to do it with a stochastic output of yes/no only (needing to guess proper outputs, and the signal you're getting isn't even deterministic / consistent). You don't even have the distribution over output tokens, you only have yes/no based on the tokens you did generate (a far weaker signal).
Re:Introducing Pangram Image 1.0
by hex4def6 on Thursday November 05, 2026 @06:40PM
(#50386178)
At least on gemini generated images vs real ones, it works surprisingly well.
Taking a photo of an AI generated image on my grubby laptop screen with my phone, moire and all, still got detected. So the signal it's using is pretty robust.
Re:Introducing Pangram Image 1.0
by sean_pedersen on Thursday November 05, 2026 @07:35PM
(#50386628)
An ill advised undertaking IMO: Any attempt to build AI generated content detection systems is flawed, since the outputs of such a system may be used to train an even better fake data generator. This leads to an equilibrium state of digital uncertainty: nothing in the digital realm can be deemed as real anymore - only as digital.
Re:Introducing Pangram Image 1.0
by karahime on Thursday November 05, 2026 @08:52PM
(#50387158)
Adding to this, the most motivated actors to get around it or find an edge case, e.g. in politics, will use it to say "Look, of course the scene depicted in this compromising image actually happened, the detector said so".
Re:Introducing Pangram Image 1.0
by 80fj0348i on Thursday November 05, 2026 @10:16PM
(#50387736)
> the outputs of such a system may be used to train an even better fake data generator.
[citation needed] - my understanding of how entropy works in real photos vs. synthetic images (AI or not) is that this is extremely difficult without severe degradation of image quality.
Re:Introducing Pangram Image 1.0
by rtpg on Thursday November 05, 2026 @10:37PM
(#50387849)
there's some truth to what one says but after hearing how Pangram works I'm pretty impressed at the methodology
I think the founder once said something like "well you can try to fight our detection and do a bunch of stuff, you just might find yourself with a pile of extremely garbled text that is less than useless".
Re:Introducing Pangram Image 1.0
by modeless on Thursday November 05, 2026 @11:05PM
(#50388029)
AI detectors can't work in theory, but Pangram works well in practice. This is because the AI companies are not trying to fool Pangram. They have no incentive to; in fact the incentive goes the opposite way. They want to watermark their outputs to leave no doubt as to whether something was generated by their AI.
Re:Introducing Pangram Image 1.0
by vintermann on Friday November 06, 2026 @05:49AM
(#50390735)
You can't trick it with AI generated text, but can you trick it into calling something AI generated which isn't? From what I've seen these AI detection tools are very good at detecting AI generation, they're also very good at calling anything AI generated.
Re:Introducing Pangram Image 1.0
by mjdv on Friday November 06, 2026 @07:06AM
(#50391398)
Pangram boasts an extremely low false positive rate, in contrast to other AI detection tools. This has matched my limited experience.
If you have text that you can prove is not AI-generated (e.g. it is demonstrably older than about 5 years) that gets an AI flag, you could try tweeting at the Pangram founder, I've seen screenshots of him offering bounties (admittedly mostly as a dare to claims he doesn't believe).
Re:Introducing Pangram Image 1.0
by gunalx on Friday November 06, 2026 @09:27AM
(#50392675)
Yes. Even as little as a "humanize and rephrast this to be more natural" prompt with high temperature on a uncommon model reduces the confidence greatly.
Re:Introducing Pangram Image 1.0
by orbital-decay on Friday November 06, 2026 @09:18AM
(#50392563)
AI companies might have no incentive but determined actors do, by definition. Anything Pangram could possibly do (even semantic variation detection) can be trivially bypassed by doing img2img on a custom trained local model, which is in itself done trivially in practice by many people. Same for text. This gives you false sense of reliable detection and false negatives as a result.
Re:Introducing Pangram Image 1.0
by monster_truck on Friday November 06, 2026 @01:58AM
(#50389067)
I'm not all that worried about this, I do have a very small concern about there only being one gig in town doing this. Creates a situation where they could be co-opted to label legitimate images as fake.
Re:Introducing Pangram Image 1.0
by program_whiz on Friday November 06, 2026 @06:28AM
(#50391048)
In theory you are correct, in practice it works. In theory you could use it as a training signal, in practice you cannot. In theory, theory and practice are the same, in practice they are not.
Re:Introducing Pangram Image 1.0
by bemtevi on Friday November 06, 2026 @07:22AM
(#50391521)
One thing I've observed is that people who use AI rarely want to hide that they're using it. With text or even images, it's sort of like a dialect. AI-generated/assisted stuff is becoming somewhat like a fashion trend where people tend to use it and it becomes part of them, rather than something they want to pass off as entirely human-made.
So I think there's very little incentive for an arms race to occur in this case.
Re:Introducing Pangram Image 1.0
by gunalx on Friday November 06, 2026 @09:29AM
(#50392714)
I dunno. Personally I like it to boost my self confidence in the text I write by doing a pass, but I really hate the overly verbose and unopinioned generic ai voice it always reflavours into.
Re:Introducing Pangram Image 1.0
by Trung0246 on Thursday November 05, 2026 @08:43PM
(#50387098)
Interesting, it detected NovelAI stuff now. Looks like one can easily defeat it by having pixel-level modifications to make AI images to have characteristics of digital brush strokes and reducing "noise" in output.
Re:Introducing Pangram Image 1.0
by omoikane on Thursday November 05, 2026 @11:11PM
(#50388082)
> An example of a real video (left) and its synthetic mirror (right).
This page would load a bit faster if the ~49MB animated GIF is replaced by a ~5MB video. I wonder if it was done this way to make sure that the video plays automatically?
Re:Introducing Pangram Image 1.0
by nwhnwh on Thursday November 05, 2026 @11:44PM
(#50388322)
The cookie modal is disgusting... It takes a few moments to realize that if you only want the necessary ones, you need to close the modal as the existing action button is for allowing everything.
Re:Introducing Pangram Image 1.0
by raincole on Friday November 06, 2026 @12:24AM
(#50388545)
From my test it works surprisingly well. I have a digital painting that is only posted very recently (therefore not in the training set). I asked GPT, Gemini and Riverflow to generate images in the same style, and use SD to add noises to remove SynthID.
Pangram gave correct results for all of them (the original one is No AI; all the others are Likely AI).
Re:Introducing Pangram Image 1.0
by sashank_1509 on Friday November 06, 2026 @04:18AM
(#50390048)
We can do this for code too I’m guessing?
I think for small snippets and well defined codebases, we probably can’t,
But if we say a fully vibecoded project vs handmade, I suspect there are AI tells outside of basic stuff like bar names etc that a model can learn, and it will be fun to know.
We could probably also do these for maths.
I guess what I’m getting at, is even in these days extremely verifiable fields, I guess AI has diverged enough that I think we can train a detector to separate AI from how a human thinks?
Re:Introducing Pangram Image 1.0
by oakashes on Friday November 06, 2026 @01:26PM
(#50396241)
There are different incentives for AI detection with text vs. images. AI detection on text is just about who gets credit for something and how much, not about the inherent value of the text itself, because the mere existence of a piece of text is never taken as proof of anything. Images are often used as proof, and so there's much more of an incentive for users of AI image generation to engage in an arms race with detection methods.
I would love if Meta licensed this tech to use for everything on Facebook, Instagram, and Whatsapp. Including ads!
Do you believe it generally doesn't work or do you believe that if you work at it you can evade them and therefore there's no point to it?
But I agree it’s good for longer text of 1000+ words.
Do you mind providing that example?
1. Social media moderation. You want to limit the proliferation of this stuff because it annoys users.
2. Trades where authenticity matters, like journalism, law, education. You have to weed out the crap and people will try to deceive you.
Think of the resources it takes to train good text generators, with perfect knowledge of the next token to generate. Now imagine to trying to do it with a stochastic output of yes/no only (needing to guess proper outputs, and the signal you're getting isn't even deterministic / consistent). You don't even have the distribution over output tokens, you only have yes/no based on the tokens you did generate (a far weaker signal).
Taking a photo of an AI generated image on my grubby laptop screen with my phone, moire and all, still got detected. So the signal it's using is pretty robust.
Where’s the line? If you take a selfie in Times Square, and there’s an AI-generated image in an ad behind you, that photo you took isn’t AI-generated.
[citation needed] - my understanding of how entropy works in real photos vs. synthetic images (AI or not) is that this is extremely difficult without severe degradation of image quality.
I think the founder once said something like "well you can try to fight our detection and do a bunch of stuff, you just might find yourself with a pile of extremely garbled text that is less than useless".
https://youtu.be/ntyZwokAfvU?si=NeA2rvBJVHAJ-qSS
1. Create AI generated text with 2+ models.
2. Merge the text (have a model just choose 'best' text). this step is to get "good" writing.
3. Use an open model to 'flatten' the text. Open models are not as good at writing, so the writing gets slightly worse.
4. Have a locally trained model on your writing samples rewrites with your voice.
If you have text that you can prove is not AI-generated (e.g. it is demonstrably older than about 5 years) that gets an AI flag, you could try tweeting at the Pangram founder, I've seen screenshots of him offering bounties (admittedly mostly as a dare to claims he doesn't believe).
Sometimes it pays to attempt the impossible.
But by analogy compression is also impossible. And yet zip moves...
So I think there's very little incentive for an arms race to occur in this case.
This page would load a bit faster if the ~49MB animated GIF is replaced by a ~5MB video. I wonder if it was done this way to make sure that the video plays automatically?
Pangram gave correct results for all of them (the original one is No AI; all the others are Likely AI).
I think for small snippets and well defined codebases, we probably can’t,
But if we say a fully vibecoded project vs handmade, I suspect there are AI tells outside of basic stuff like bar names etc that a model can learn, and it will be fun to know.
We could probably also do these for maths.
I guess what I’m getting at, is even in these days extremely verifiable fields, I guess AI has diverged enough that I think we can train a detector to separate AI from how a human thinks?
I’m honestly surprised pangram got this right, with my human eyes it feels indistinguishable but apparently not !