Scammers are using AI to lure shoppers to fake businesses (www.bbc.co.uk)

Fraudulent overseas sellers are increasingly using AI‑generated images and fabricated backstories to pose as UK family-run boutiques, luring shoppers with targeted social‑media ads and charging premium prices for cheap, mass‑shipped goods. The BBC and consumer group Which? highlighted fake stores such as “C’est La Vie” and “Mabel & Daisy,” which presented believable husband‑and‑wife or mother‑and‑daughter personas and UK addresses while using returns addresses in China; hundreds of one‑star Trustpilot reviews describe poor quality items, extortionate return fees and long delays. Regulators including the ASA have already banned some deceptive ads, but under‑resourced trading standards and platforms’ ad ecosystems mean many scams go unchecked. For the AI/ML community this is a real‑world case of synthetic media enabling large‑scale social engineering and reputation laundering. Experts note visual giveaways (overly “perfect” staged images, earlier failure modes such as unrealistic hands) and recommend detection heuristics like consistency checks across images, different backgrounds, and verifiable location metadata. Longer‑term mitigations include provenance/watermarking for generated imagery, improved platform moderation, automated synthetic‑content detectors, and tools to surface supply‑chain signals (payment/return addresses). As generative models improve, the challenge will shift from spotting AI artifacts to proving the presence of any real human operators — raising urgent technical and policy questions about trust and verification online.