Woman created deepfake videos to force rivals off daughter’s cheerleading squad | Hacker News Reader

Woman created deepfake videos to force rivals off daughter’s cheerleading squad

inquirer.com

105 pointscoloneltcb75 comments

From TFA (and making assumptions about the woman), I don't believe these were "deepfakes". It sounds like it was just the usual Photoshop kind of stuff. And in that regard, it's just another sad story about some people with issues.

It's hard for the general public to grasp the difference between a photoshopped image and one generated with ML processes, so I expect both will be called "deep fakes" by the media for the foreseeable future.

That’s unfortunate. But also probably true.

I think this quote suggests you are right:

> The videos were analyzed, and detectives were able to determine they were “deepfakes” — digitally altered but realistic-looking images — created by mapping the girls’ social media photos onto other images...

Both this article (and the Inquirer article it is based and basically plagiarizes - https://www.inquirer.com/news/bucks-county-raffaela-spone-cy...) seem to be confused about the difference between photos and videos and seem to imply that they are interchangeable. It would be interesting to read the actual affidavit and see if it contains the same confusion, or if this is something caused by the Inquirer's reporting.

Ok, I've changed the URL from https://www.pennlive.com/news/2021/03/pa-woman-created-deepf..., in keeping with the HN guidelines' request for original sources. Thanks!

https://news.ycombinator.com/newsguidelines.html

> the Inquirer article it is based and basically plagiarizes

Seems hard to accuse it of plagiarizing the article when it cites that article as its source three separate times, in the only three links that appear.

The PennLive article is an example of "close paraphrasing" [1] or improper source citation. There are several sentences and phrases in it that are extremely similar or the same as those in the original Inquirer article. This is even true in the lead sentence - both articles contain the same phrase to describe the "deepfakes": "fake photos and videos that depicted the girl’s rivals naked, drinking, or smoking".

[1]: https://en.wikipedia.org/wiki/Wikipedia:Close_paraphrasing

That's not improper source citation. That's how newspapers work.

Wait until you hear about the Associated Press.

She was dumb enough to upload the photos directly, exposing her IP address. I doubt she's smart enough to make deepfakes. This is most definitely a Photoshop story.

There are plenty of apps out there that create image and video deepfakes with just a few clicks, with impressive results.

Does the distinction matter though? It's getting easier every day for people to access commodity tools to make deep fakes, and it's only a matter of time until the cambrian explosion of indistinguishable visual media arrives.

It did mention videos. That makes me think they're deep fakes.

I thought so too, but then I re-read the article and the original report linked from the article, and they both seem to use “video” and “photo” interchangeably. Whenever specifics are given, though, it’s always a photo. Despite the headline, the reports provide no evidence that there were actual videos involved, AFAICT.

Maybe it was a video slideshow.

There aren't any videos mentioned in quotes, which makes me think that it's likely a mistake on the reporter's part. This paragraph seems to conflate the two.

> The girl and her coaches were sent photos that appeared to depict her naked, drinking and smoking a vape pen

"Smoking a vape pen" I don't get this one. Was it an offence to be a cheerleader and vape? Why not go all out a do a cigarette, they were already going all out with naked photos.

I'd guess some of the photos were real, but the police are calling them 'deep fakes' for privacy reasons.

In some areas, vaping is seen as something only the dregs of society would do. In other places, it's just seen as an alternative to smoking. But smoking itself is seen as a low class or high class activity as well depending on where you are.

Also vape usage trends young, so depending on who the coach is they may have a more negative view of it if they don’t have any peers doing it, and they are supposed to be discouraging the kids from doing it. Unfamiliarity breeds distrust.

> the usual Photoshop kind of stuff

https://news.ycombinator.com/item?id=26448986

Here is a great PSA on how to defend yourself against DeepFakes https://www.youtube.com/watch?v=_r-r6lFheKA

That video just says to listen to the voice, but faking the voice is also part of the technology now

Also, people exist who are good at voice mimicry. I can do a number of celebrity voices and some of my friends too.

Which itself was entirely faked.

Nice that the got the criminal but considering we're entering the "post-truth" era... how can we check this news is true?

Not wanting to fuel any conspiracy theory, just genuinely curious.

Same as always, by considering available evidence, gathering testimonials, and determining the most likely explanation. The increasing abundance of recorded data in the custody of a growing number of entities likely outcompetes the ability to fabricate evidence, making objective truth more knowable than ever.

So then attempting to live in privacy by reducing tracking and data recording becomes a liability. Great.

I envision a possible future where people wear tracking and recording technology at all times to provide plausible counter-narratives to deep fakes.

I hope we just do the rational thing, and rather stop treating video as if its some perfect representation of reality that's inherently trustworthy.

We've had to confront similar issues with other media earlier in history, after all.

Yet without video many terrible crimes would go unpunished. What else do you propose?

Cryptographic chain of custody from CCD to MPG, with public keys for video recording HW burned in at the factory. I don't know if that's at all feasible. Possibly only for security cameras.

No, god damn it. As soon as the cat's out of the bag, it'll forever be an arms race (there'll be attack vectors, etc.).

Let's not just establish a new proxy for implicit trust here.

What qualifies as evidence, then? Trust is important, but it can only exist long-term if there's a way to detect lies/betrayal. Anyway, I expect market forces to steamroll both of our opinions.

I don't know – but as far as I see it, that's irrelevant in pointing out that video (in and of itself) should not be trusted by virtue of being video.

The fact that you can't imagine – and I don't provide – any alternative proposition does not make implicitly trusting video a good proposition.

But, yeah, I suspect reality will probably disappoint us both...

Uh, right, implicitly trusting video because it's video is definitely dumb. I don't think that's what I'm proposing. Trusting cryptography, as long as you're explicit about what you're doing and how it can fail, is at least not immediately foolish.

It has nothing to do with the medium of video, but with establishing trust.

We'll need to deal with video the same way we do with images, audio, and eye witness accounts. The way to do that is to establish a measure of trust in some quantifiable or methodological way, rather than just trusting the medium.

Google Glass tried that. Even with an obvious indicator of recording, some people still believed that anyone wearing Google Glass was recording 24/7. There's no way that'd fly in modern society.

Recording indicators are doomed to fail, as people want a not-recording indicator. The only obvious and reliable one is having the camera out of sight.

In any downtown area you are on camera 24/7. I think the expectations of privacy have changed. Google glasses released on mass now would not make similiar headlines.

Go to a public area and visibly record people. Particularly while talking to them. You will annoy a lot of people and some will act with hostility (especially if you don't frame it as some sort of interview).

People don't like it when nobodies are filming them, even in very public spaces. For some reason they are ok with CCTV, despite that not being much different (I do recognize that cctv is typically recording a general area and isn't always HQ)

I think people don't mind being recorded but they want to know why. If the reason becomes because a big name company does it automatically we generally accept. For example people now accept amazon / google listening to background conversations in order to process commands.

'downtown areas' have been on camera for decades, certainly by the time google glass was released, so i don't think this is really a useful data point.

I agree. I wouldn’t at all be surprised if “civilian” bodycams become the norm in public in the next ten years. It’s just a matter of making them convenient and fashionable enough.

We already do. It's called a cell phone.

But at least my cellphone doesn't track my every movement and my surroundings to enable me to defend against unfounded accusations and deepfakes, it does so to help Google sell ads.

Not looking forward to the next year or so, when consumer-grade tech is good enough that the average detective won't be able to detect fakes so easily.

To the contrary, given that the deepfake cat is out of the bag, I would rather deepfakes become normalized such that people stop putting so much faith in video evidence.

I'm worried about State-grade tech...and with that political propaganda power.

Propaganda works already perfectly without deepfakes.

There is never a perfect, a Picture or a Video say's more than thousand words. But yeah, your right it's already honed to near perfection.

s/deepfake videos/photoshopped images

How do you go about gathering data and training a model for hundreds of hours and not question that what you're doing is wrong?

> How do you go about gathering data and training a model for hundreds of hours and not question that what you're doing is wrong?

Technical competence is not a proxy for ethical behavior.

You know, you could ask a sizeable chunk of SV the exact same question.

Even if the images represented reality, what the mother was doing was still harassment, particularly with the nudity angle.

She could be facing much more serious charges for that too. It sounds like these girls are minors.

She's preparing the daughter for med school or law school, and maybe a career with an MBB.

Daughter's got to learn the Six Sigmas somehow:

1. Teamwork

2. Insight

3. Brutality

4. Male enhancement

5. Hand-shakefulness (post-COVID)

6. Play-hard

Devil's advocate: Without getting caught, one has to ponder the malice vs. the expedient and rational considerations for competitive genetic advantages conferred, socially and scholastically.

A lot of people are considering what it might look like to live in a “post-truth” world. To me events like this suggest projects like ChainLink and the Band Protocol might become vital infrastructure of truth much sooner than many think.

Maybe forensics/criminology is already there but I wonder if detailed study of fake videos will become a very in/demand skill / specialization.

Lawsuit filed in 2014. Deepfake then, not so much.

How did she get caught?

You might be interested in this news story which goes into the details of how she got caught: https://www.pennlive.com/news/2021/03/pa-woman-created-deepf...

You know, I read the article, found your comment, upvoted it because I would have been interested in the details of how she got caught, and then saw it's the same article. This does not go into the details of how she got caught; that is dealt with only in this single sentence:

I'm glad we can undo upvotes now, I guess.

The news article links to:

https://www.inquirer.com/news/bucks-county-raffaela-spone-cy...

Which IMO does a much better job of explaining how she got caught.

Literally the paragraph after the one you quoted:

> Officers were able to execute multiple search warrants that allowed them to trace the text messages back to Spone’s IP address and then her cell phone, police said.

That's a question of how they attributed the images to her. It's not a question of how they determined that she faked the images.

Username checks out.

Raffeala Spone, the former cia black propaganda operator and specialist...on a new and different mission, watch season one now!!! A netflix original :)

There’s a grain of truth in there somewhere.

Reply on news.ycombinator.com