Searchable transcript of 4 AI Scams Targeting Ecommerce Sellers (We Got Hit By One) — MyWifeQuitHerJob Ecommerce Channel (14:21). Search for a phrase, then click its timestamp to jump straight to that moment in the video.
Captions sourced from the original video on YouTube, published by MyWifeQuitHerJob Ecommerce Channel. The video, its captions and all related intellectual property remain the property of their respective owners; AINotes claims no ownership. Provided for research, accessibility and search — see the Transcript Notice and Copyright Policy.
00:00 So, I caught a shopper trying to scam my store last month. AI is now good enough that customers can generate fake photos, fake shipping labels, fake chat histories, and then use them to steal thousands of dollars from you without ever leaving their couch. And here's the part that pisses me off. I caught this scammer red-handed, but I still ended up paying them.
00:21 Because in e-commerce right now, fighting the customer often costs more than eating the loss. Here are four AI-powered scams that you need to watch out for as a seller. So, let me start with the scam that I got hit with in my own store, because it is the most common one going around right now. I run a seven-figure store that sells handkerchiefs and linens.
00:40 We do customer embroidery on almost everything. And because we are a small operation, we pack every single order ourselves. There may be four of us in the warehouse on any given day. So, if a customer complains that the embroidery on their handkerchief looks bad, I can usually just walk over to whichever employee packed that order, and then ask them about it.
01:01 We remember these things because our products go through an entire hand-finishing and QA process. Every piece gets stitched, ironed, and the threads trimmed by an actual person. So, last month, a customer sent us a photo of a handkerchief with horrible-looking embroidery, asking for a refund. My team pulled the order, and we all remember that specific piece going out clean.
01:23 The embroidery in the photo had to be AI-generated. Somebody had taken a real picture of our product, and then had an AI model add sloppy stitching to it. And here's what really got me. Even though we caught the scam, we still sent them a free replacement. Because in my store, arguing with the customer over a $30 handkerchief costs more in time and headache than just eating the loss.
01:45 The scammer wins either way. Now, if you think this is just a problem for small stores, let me tell you what happened to Boll & Branch. They are a luxury bedding brand that does hundreds of millions of dollars in revenue. Their CEO, Scott Tannen, posted about this on LinkedIn. A customer submitted photos of torn sheets asking for a refund. But the tear did not look right in the photo because cotton doesn't fray that way.
02:10 And on one of the photos, the customer had even forgotten to remove the AI watermark before submitting it. So instead of giving a refund, they asked the customer to hop on a FaceTime call to verify the damage in real-time, and then the customer never responded. But here's the scary part. Tannen had his team do a complete audit of previous return tickets, and they found hundreds of photos that appeared to be AI-generated that had already been refunded.
02:34 So it is scary what AI can now fake. It can generate cracks, stains, leaks, mold, dents, tears, missing parts, or any other damage a customer could ever claim about a product in about 30 seconds. Now there's a subreddit called Is This AI? that has over 400,000 members whose entire purpose is to ask whether random product photos are real or generated.
02:55 Now if 400,000 normal people can't tell the difference, well your customer service team won't be able to, either. But this photo scam is just the tip of the iceberg. And if you thought the photo scam was bad, this shipping scam is way worse. So here's how it works. A customer buys something from you, and when it arrives, they tell your customer service team that they want to return it.
03:19 Your rep sends them a return label, and then the customer uses AI to generate a fake shipping receipt from UPS, FedEx, or USPS. They send that screenshot to your rep, and they say, "Hey, I dropped it off, but the tracking hasn't updated yet. Can you refund me right now?" But you don't want to argue because you're terrified that the customer will leave you a bad review.
03:41 You see what looks like proof of a drop-off, so you click refund, and the package never actually gets returned. Modern Retail broke this story recently. The AI-generated shipping receipts are getting good enough now that humans cannot tell them apart from real ones. But there is a version of this scam that is even more sophisticated. And it is called the fake ID tracking scam, and it works like this.
04:05 The scammer uses AI to change the return address of a store-issued return label to a fake address. And then they ship an empty box, and that empty box gets delivered to this fake address. Your tracking system sees the package as delivered, and then you see the delivery confirmation, and then the refund gets issued. Now, the scammer has both the original product and their money back.
04:30 Now, personally, all this terrifies me because our default policy in our store is that the customer is always right. If somebody claims their package never arrived, and they seem honest about it, we just refund them. We don't argue with our customers, and honest mistakes happen all the time. Now, the real problem shows up when there are entire groups of people organized specifically to run this scam over and over, ordering from you every week, and then turning your good faith return policy into their own personal ATM.
04:59 So, recently, we made two changes to our operation to defend against the organized version of this. So, first, we started taking photos of large orders before shipment. If a customer claims they never receive something, we can just pull the photo and the shipping weight, and then cross-reference it against what was supposed to be in the box. This is a pain in the butt, but it has saved us from a lot of fraudulent claims.
05:23 Second, we started offering shipping insurance at checkout. Customers pay a small fee, and if their package gets lost or never arrives, we refund them, no questions asked. Now, scammers usually skip this because paying for the insurance eats into whatever they were planning to steal, and this money also acts like a reserve fund to cover the fraud that makes it through.
05:43 But you have to be careful not not overdo it because you can't really crack down hard on these scammers. The National Retail Federation found that 82% of consumers say free returns are a major factor in where they shop. And 76% say that they want an instant refund option. So, the more verification you demand, the more legit customers you push away. Whereas, the less verification you demand, the more scammers eat you alive.
06:10 And here's another thing to consider. None of these scams would work if sellers were willing to fight back. But, sellers are terrified of getting a bad review. On Amazon, a single bad review can wreck your ranking and cost you tens of thousands of dollars in lost sales. On your own store, a public complaint can spread on social media in a single day.
06:30 So, every seller in this industry has been trained to just eat the loss and issue the refund. And the scammers know this, and that is exactly why this fraud works. Now, up to this point, every scam I've shown you has one thing in common. A human being on the scammer's side still has to make the request. But, these next two scams do not even need that, because AI is running the entire attack at scale 24 hours a day.
06:56 By the way, if you're interested in learning how to start your own e-commerce store, make sure you sign up for my free 6-day e-commerce mini course below. It's 100% free, and I guarantee you'll learn a ton. So, scam number three is an army of AI voice bots attacking your customer service line. And here's what this scam looks like. A scammer buys a bunch of stuff from an online seller.
07:17 They then load an AI voice bot with the real order numbers and whatever partial customer info they have from a data breach. The bot then calls your customer service line sounding like a real human customer asking for a refund. And when your rep asks for verification, the bot answers them because it has real data and the refund gets issued. Now, unlike the other scams, which are just one-off calls, these bots are calling your line hundreds or thousands of times a day.
07:44 Pindrop, which is a deepfake detection company, put out a report saying that three in 10 retail fraud attempts are now AI-generated. And some large chains are reporting more than 1,000 AI bot calls per day hitting their customer service lines. Now, let me give you a specific example that really disturbed me when I read it. One retailer discovered that AI bots were calling in requesting $21 refunds across thousands of calls.
08:09 Every single request was for exactly $21. Now, why is that? Because $21 was just under that company's threshold for requiring a manager approval on a refund. So, every one of those refunds went out without a single question being asked. So, this is fraud at scale run by organized operations. They purposely buy a whole bunch of stuff from you and then unleash an army of AI bots to extract the refunds and free merchandise over and over again 24 hours a day.
08:40 Now, you might be thinking to yourself, "Come on, these schemes are so obvious. Anyone can detect this and put a stop to it, right?" Well, let me tell you the DoorDash story. A woman in Wisconsin ordered chicken wings, ate part of one, took a photo, and then used ChatGPT to edit the photo so that the wing looked raw. She sent it to DoorDash and they refunded her almost $40 in credit, no questions asked.
09:01 And then, she posted a TikTok bragging about the whole thing and it got 4.4 million views. Now, the only reason this woman got caught is because she publicly bragged about it to 4 million people on TikTok. DoorDash does not have a solution to this yet and their only line of defense is just flagging high-risk customers. Now, this final scam is one that uses humans to attack AI customer service bots.
09:30 So, if you're a seller who's already using or considering an AI chatbot to handle customer service, you should pay close attention to this one. Because the legal precedent has already been set and it is not in your favor. So, there's this guy named Jake Moffatt. In 2022, his grandmother died and he needed to fly from Vancouver to Toronto for her funeral.
09:52 So, he went to Air Canada's website and asked their chatbot about bereavement fares. And the chatbot confidently told him that he could book a full price ticket and then apply for a bereavement discount retroactively within 90 days. But this policy didn't actually exist. The chatbot hallucinated it and then Jake booked the flight based on what the chatbot told him.
10:14 And when he tried to get the discount, Air Canada refused, so he took them to court. Air Canada's defense was that the chatbot was, quote, a separate legal entity responsible for its own actions. But Air Canada lost the case. The court ruled that Air Canada was responsible for everything its chatbot said because the chatbot is an extension of the company.
10:36 Air Canada had to pay Jake back his money plus court costs. So, basically, if your bot says it, you own it now. Now, let me show you the scam. A Chevrolet dealership had an AI chatbot on their website to help with customer questions. A user prompt engineered this bot into agreeing to sell a brand new $76,000 Chevy Tahoe for just a buck. And because of the Air Canada precedent, the dealership had to fight the sale.
11:04 So, stuff like this is happening all over the place. A UK retailer's after-hours chatbot was talked into offering an 80% discount plus a full refund on an 8,000-lb order. And incidentally, that is why I have yet to deploy an AI bot on my site. The whole point of a chatbot is that it's convenient and responds 24/7 to whatever a customer types in. But you literally cannot train a chatbot to refuse every possible prompt because smart users can manipulate the conversation.
11:34 So, if you're a seller who's already deployed an AI customer service bot, or you're thinking about deploying one to save money on your support team, understand what you are actually signing up for. Because you are creating a new target for scammers to exploit around the clock, and the courts have already ruled that whatever your bot promises, you have to honor.
11:53 Now, all these scams sound scary, so what can you do as a seller? Well, if you're running your own store, you have the most leverage. So, first thing you got to do, whenever possible, require the customer to physically mail the item back to you before you issue a refund. It's a pain in the butt for both sides, but it dramatically cuts the fraud rate.
12:13 Happy Returns data shows that in-person verification alone reduces return fraud by at least 85%. Second, on any damage claim, ask for multiple photo angles, or even a short video of the actual product. Video is much harder to fake than a single photo, and asking for it filters out most casual scammers. Third, take photos of every single outbound order and log the weight.
12:38 If somebody claims that they never got their order, or that something was missing, you can pull the photo, and then cross-reference it. And this has saved us thousands of dollars. Fourth, offer shipping insurance at checkout. Legit customers who worry about losing a package will pay the small fee. Scammers will skip it because it eats into their fraud margin.
12:59 But, unfortunately, if you're selling on Amazon, you can't really do anything. Your best move is to use Amazon's returnless resolution feature. As of November of 2025, Amazon lets you set a price threshold between $1 and $75 for automatic refunds without requiring the customer to return the item. So, for low-value items, where the cost of processing a return exceeds the loss, turn this feature on.
13:25 But, make it harder for high value returns. But ultimately, here's the reality. That handkerchief customer who scammed me last month is probably still ordering from other sellers today, running the exact same playbook. And every one of those sellers is doing what I did. Catching the scam, knowing it's a scam, and refunding them anyway. Because in e-commerce right now, fighting a bad review costs more than eating the fraud.
13:50 Now, our store hasn't been crushed by this yet, and I hope it stays that way. But every seller I talk to is starting to see this. And on Amazon, you don't even get the option to fight back. FBA process the return, the refund goes out, and you find out about the fraud on your balance sheet. Which is just one more reason to build your business off of Amazon while you still can.
14:11 And now that you understand what's going on, make sure you watch this video here on exactly what I would do if I were to start an e-commerce store from scratch today.