Red Flag Academy

Let’s learn how to spot scams and misinformation.

[redflagacademy]

Recently on my main Tumblr account, @isitsafe, I posted a viral AI video with the intention of pointing out the telltale signs of AI-generated video. As generative AI becomes more and more ubiquitous, and significantly more convincing, I feel learning how to distinguish the real from fake is imperative.

As of right now, I’m not sure how I’m going to go forward with this project, but I do know I want to use this Tumblr account to discuss misleading AI, internet scams, and maybe dodgy Kickstarter campaigns. Stuff like that.

So, with that in mind… welcome to Red Flag Academy.

[redflagacademy]

I’m only posting this because someone might see it and panic.

Amazon was not hacked*. She was hacked. Most likely she inadvertently got phished (sent an official-looking email or text with a link designed to get you to enter your login information on a fraudulent site).

Amazon rarely sends emails out that require immediate action. Don’t click any weird links.

Use your browser’s random password generator. Let go of the idea that you need to remember your passwords. And definitely stop using the same password on multiple sites.

Make sure 2-factor authentication (2FA) is on. If you’re one of those people who thinks 2FA is inconvenient, stop being that person. If your login credentials are stolen, 2FA is the first line of defense. The most convenient method is to have a code sent via text. You’ll be immediately notified if someone tries to sign in to your account, and you don’t need a smartphone. Authentication apps are slightly faster and a bit more reliable, but they of course require a smartphone.

*There was a recent breach of the European Commission database, but that would not have made her account vulnerable in this way.

[redflagacademy]

I’ll keep this one brief. Tap where it says “Joined”.

Our POTUS has no authority over South Asia.

@/NameRandomnumbers is engagement farming.

In these examples, foreign accounts are disguised as being from the United States. The content they post is almost all catered to U.S. tastes.

And they follow and repost each other to generate revenue through Twitter’s blue check program. It’s usually a harmless scam (but indeed a scam) until they start posting disinformation. I suggest blocking such accounts.

[redflagacademy]

I should warn you, this video is disgusting and a bit disturbing.

Thankfully, it’s also fake.

I don’t think I need to point out all the obvious AI red flags. With this particular video and story, let’s look at the bits that defy common sense.

No (reputable) source - If something like this was really happening, thongkhanmk789 wouldn’t be the one breaking the news to you.

No litigation - Remember when McD’s served coffee so hot that it severely burned a customer? She sued and won. If multiple people were somehow getting severely sick or even having seizures from their ice cream the lawsuits would be all over the news.

No way - The chances of “coliform bacteria” (for some reason shown as maggots and mold in the video) getting into their ice cream machines is low. It’s more likely to find it in meat, vegetables, and ice from an improperly cleaned dispenser, all of which have occurred at McDonald’s and made headlines at the time.

Also, the ice cream machines are broken.

[redflagacademy]

Right here, @kangamommynow.

AI Image Detector - Identify AI-Generated & Manipulated Images | TruthScan

Advanced AI image detection technology. Identify deepfakes, AI-generated images, and manipulated photos with 99%+ accuracy. Trusted by enter

[redflagacademy]

My shortest post.

[redflagacademy]

🚩Amazon does not text customers messages like this.

🚩This text was sent to 9 phone numbers, even though there’s only one order number.

🚩The texts they send come from a five digit number.

🚩Amazon FAMOUSLY has its own website and domain.

[redflagacademy]

Online Shopping (Amazon)

Everyone shops at Amazon, and that’s barely an exaggeration. They sell pretty much everything you can legally purchase. This is because pretty much anyone can become an Amazon seller.

The biggest red flag is unknown and nonsensical brand names. The second, AI-generated or manipulated photos of the product. It is not hard to find examples.

I’m just saying, someone at Amazon might see these images on their screen and compared them to the actual product in their hands and thought, “Somebody’s about to be disappointed.”

Sometimes the sellers go absolutely nuts with Photoshop.

There’s no real image of the product. You might think it’s weird that the seller couldn’t at least sit product on a desk and snap a photo with their phone or something. Instead you get CGI renders like these below.

Oh, yeah. That’s the third biggest red flag. It’s called white labeling. White-label products are made by one manufacturer, but are rebranded and sold by multiple retailers. That’s the same product in the three pics above, but from three different sellers.

These red flags are all over Amazon, to be quite honest. But luckily, returns are pretty easy. But to save time, buy from brands you trust, and prioritize products shipped directly by Amazon. Look under the “Add to cart” button.

Tapping the name next to “sold by” (in this case another nonsensical name) will lead you to a page with customer ratings. I’d say it’s best to pay more attention to the negative reviews. They tend to be more honest.

[zagrebinaa]

[redflagacademy]

Well, @kangamommynow , I have some good news for you. These are definitely real photos of a definitely real place, the Palazzo Lana in Italy. There are more photos and information in the link below.

On sparkling form

The birthplace of an acclaimed Italian fizz, the Palazzo Lana in Lombardy is very easy on the palate, with elaborate frescoes and friezes ga

It’s hard to believe sometimes that there is still beauty in the world. But this photo actually has none of the typical red flags of an AI-generated image. Here are some green flags though.

The light and shadows are natural and consistent.

The artwork in the background and at odd angles at the sides look solidly structured.

The two blue chairs match without being identical.

A reverse image search leads to a reputable source that includes the photographer in the byline. (Paola Pansini, link to her website below.)

PAOLA PANSINI

PAOLA PANSINI

[redflagacademy]

Political Rage Bait

I’ve said before that rage bait can be fun. But like most things that are fun, it can deployed irresponsibly. Example:

This post (link to source) was likely made to create rage in people who don’t like, don’t know, or are indifferent to Jon Stewart. And judging from the responses, it was successful. It’s the combination of the caption with a quote along with poster’s comment that makes the rage bait work so well. People get mad so fast they don’t even stop to think about what they’re mad at.

So, what should you do when presented with political rage bait? First thing is to think of what’s at stake. It seems innocuous at first because neither Stewart nor the Clintons are currently active politicians. Also, it’s very obvious that the Clintons were only subpoenaed by Trump’s DOJ to distract from Trump’s own involvement with Epstein. So, you might want to look into finding out why Jon would say what the post claims he said.

Did he say it? Yes. But notice the caption. It only directly quotes one word, the most eye catching word. And there’s a source mentioned to where the poster found the quote (The Hill), but he doesn’t provide the a link to the article. That puts extra steps in between seeing this caption and finding the full context. Most people won’t bother taking those steps (but I did, link).

If you find the article, or if watch the short video clip the article is about (link) you can know the full quote.

On The Weekly Show podcast, Jon is asked if Bill and Hillary should comply with subpoenas from the House oversight committee into the investigation of Epstein, and Jon answered:

I absolutely do. But why should they comply if the Department of Justice is not complying with releasing the files? Like… Is compliance a kind of specialized, individual, indigenous opportunity, or is it… should it be universal? The Department of Justice has subpoenaed them to testify in the Jeffrey Epstein case, while not complying with releasing the files. So how does that comport in any? But do I personally think they should comply? Abso-fucking-lutely. And if they’ve got something to hide or nefarious, yes, we should know about all this. It’s bonkers how long this is going on.

But you can’t really make an eye catching, scroll stopping post with all that.

If you know me from @isitsafe, you know where I stand politically. And it’s not unreasonable in my mind to want everyone with known ties to Epstein to be investigated. My opinion of Jon Stewart would in no way change that.

You can protect yourself from irresponsible political rage bait by remembering that multiple things can be true at once, and by not engaging in whataboutism (a possible topic to be discussed here in the future).

[redflagacademy]

[redflagacademy]

I hate to do this because the video is hilarious to me. But I have doubts about its authenticity. That’s the biggest problem with generative AI existing as it does today. I can’t fully enjoy a video of ICE agents taking an L from actual ice.

Oh, well.

The video might be real, but always be skeptical. (Link to source)

What are the red flags I see?

Immediately, there’s no sound. Even when AI does great with video, it will do horribly with sound and speech. It can be a dead giveaway that a video is AI, so it makes sense to post a silent AI video, or to change audio.

I’ve mentioned short videos tend to be AI. It’s expensive on the consumer end to make longer AI videos. So a lot of the AI videos you’ll come across will be 10 seconds or shorter.

Video quality is more of a yellow flag (if that is a thing), because these days even the most inexpensive phone cameras are’s capable of taking a clear video. However many factors can affect the quality of video that can’t be helped by the camera or its user.

The biggest red flag in my opinion that has me 89% sure that this is AI have to do with the power lines. As the camera moves, the power lines appear to be flat on the ground. If you generate a video from a still image, AI will generally have trouble with depth and parallax. One item may be in front of another, but AI might determine that those items are closer than they are. Or if a bird was passing over the ground and seen from an overhead angle, AI wouldn’t really know whether the bird was flying or walking.

And there are other red flags, but I’m trying to keep these posts short. Basically, I just wanted to share this video, but I could not in good conscience present it as a genuine video.

It’s entertaining though.

[redflagacademy]

AI Catfishing: Ginny Part 2

The rest of these Ginny photos (link) are much easier to tell that they are AI-generated, even though they may seem very convincing. The days of warped faces and extra fingers are mostly behind us. If you are trying to judge if a photo is authentic, it is better to look at the scenery and background elements. Like this one…

Ginny’s target audience is not going to notice that her earrings don’t match. So they are definitely not going to notice the knives and forks on both sides of her plate. AI is rather lazy sometimes when it comes to backgrounds. This is why the table settings look weirder the further back you go.

Yeah, they would never install benches up against the lockers. Also, the logo is distorted on the back of her sneakers in s banner typical of AI.

The window is seen in the mirror’s reflection, but natural lighting coming from the other side of the room. Also, Ginny is wearing a bodysuit, although the mirror behind her shows her in a dress. Mirrors are a common pain point for AI. That’s why the angle and distance between the mirror in front of Ginny compared to the one behind her seem off.

The next one was kind of fun.

So many reflections. And judging by those reflections, there are no doors in this closet. No way in, no way out. It’s a fashion prison. #FreeGinny.

And I’ve clustered these together because all of them are supposed to be bathrooms. They don’t match, except for the fact they consist of a sink, a mirror, and nothing else. Three have evidence of having a window, and one is likely another completely sealed room. #FreeGinny.

Also, AI did a surprisingly good job of remembering to reverse the Nike Swoosh for the mirror reflections, except for one time in these photos. But al of the Swooshes don’t follow the curves of the fabrics.

What do you do if you encounter an obvious catfishing profile? Of course report them on whatever platform you find them.

[redflagacademy]

AI Catfishing

What is that? You probably know catfishing in regards to social media as when someone steals photos from a real account to make fake account to scam or bait people. Stealing photos is still the preferred thing to do for romance scammers, but AI-generated photos are big for people who just want attention (especially on monetized platforms).

Let’s check out “Ginny” on Twitter (Link).

This post for some reason showed up on my feed.

Anyone who has seen a lot of AI photos might immediately recognize this as one. It looks real, but feels off. AI is always getting better, yet it remains inconsistent.

Notice the plaid pattern around the waist, and how the pleats aren’t all reaching the bottom hem. And there’s a TV on a nightstand in front of a doorway.

If you’re not convinced, I’ll be back later.

[redflagacademy]

Why is Allrecipes posting clickbait? It is so flagrant that I refuse to click on it. In fact…

They started as an ice house in Dallas called the Scotland Ice Company in 1927. They expended and changed their name to Tote’m Stores in 1928. And in 1946 they became 7-11 reflecting their expanded operating hours of 7am to 11pm.

Did I do it? Did I blow your mind? Is your mind blown? 🤯

[isitsafe]

If you don’t know who DJ Akademiks is, good.

All you need to know about him is that he’s lying. Anthony Joshua was a passenger in that accident, not the driver.

Anthony Joshua, boxer who defeated Jake Paul, injured in car crash that killed 2 in Nigeria

The crash happened on the Lagos–Ibadan Expressway. The heavyweight — who recently trounced the social media star-turned-boxer in a much-pub

[redflagacademy]

And now, the first official “lesson” from the RFA…

Rage Bait!

Rage bait is anything posted to social media with the intention at attract engagement by making the audience upset. I post a lot of this lately on my main Tumblr because lots of things make me mad. And mad is such a fun emotion sometimes. However, there are red flags to look out for when it comes to rage bait.

Most people post rage bait to expose an injustice, or to simply vent. That’s fine. But it can be used to mislead, misinform, or even to sell something. Here’s an example I would like to discuss. (Summary below. You don’t have to watch the whole thing. This isn’t a real academy.)

This is Anne Sisteron on Instagram, with a video titled Cultural Appropriation at Louis Vuitton. It’s provocative. In fact, so far it has provoked over four million views.

It is massively more popular than her other videos, including one about helping dogs! The rage bait was successful, but where are the red flags?

Here’s the thing. The red flags in this video are hard to spot unless you know what you’re looking for. But if you learn only one thing from this let it be this; be skeptical of everything the internet people tell you.

Nothing she said was a flat out lie. But facts were misrepresented. She posted this video in 2025, but LV released the bag in question back in 2007. And the cheap bag that inspired the LV originated in Hong Kong in the 1960s, over twenty years before its association with Ghana. LV never called their product the Ghana Go-To bag. The original bag did get the nickname of the Ghana Must Go bag when over s million Ghanaian immigrants were expelled from Nigeria in 1983.

Back to Anne, she used her reinterpretation of the facts to deliver a call to action: “… vote with your dollars.” This suggests a boycott of LV’s goods.

However, depending on how you see the video, you might notice a button saying “view shop”.

And if you tap that button, you will be shown a shop to view.

And in that shop you can buy the jewelry she’s wearing in the video, along with other luxury pieces from her own shop. And at 4 million views, it’s safe to assume she made a sale or two, ironically by exploiting the very same situation she accused LV of exploiting.

So, if you already had solid knowledge of LV’s plaid laundry tote, the deportation of West Africans from Nigeria, and the origin of Hong Kong’s red-white-blue bag, you saw the red flags in Anne’s story immediately. And you should also try out for Jeopardy. The rest of us just need to be more skeptical, taking few things at face value.