Scammers are increasingly leveraging artificial intelligence to create realistic images for fraudulent purposes, making it harder for individuals to distinguish between genuine and fabricated content. These AI-generated visuals are being used to lend credibility to fake stories, solicit money, and extract personal information.
The effectiveness of AI image generators has advanced to a point where traditional methods of spotting fakes, such as looking for inconsistencies in the number of fingers or garbled text, are becoming obsolete. Modern AI models can now produce images with accurate details, rendering visual inspection unreliable as a sole means of verification.
Instead of relying on visual cues, experts advise a shift towards verifying the authenticity of an image and maintaining skepticism towards the accompanying narrative. Scammers often exploit urgency and emotional appeals to bypass critical thinking, making it crucial to question the context and origin of any image presented.
One common tactic involves fake lost pet appeals, where scammers post AI-generated images of distressed animals in online community groups. They then request rehoming fees or other payments, with the AI-generated nature of the image preventing a reverse image search from revealing an original source. Similarly, individuals searching for missing pets may be targeted with AI-generated photos and subsequently asked for a reward or deposit.
Dating profiles are another area where AI-generated images are prevalent, presenting individuals with seemingly flawless and consistent pictures. While video calls can be used, even these are not foolproof, as real-time deepfake technology can pass basic tests. Researchers suggest requesting unscripted actions during video calls, such as turning one's head or picking up a random object, and being wary of those who refuse to engage in a video call altogether.
Artists are also falling victim to AI-generated portfolios presented as original work on platforms like X, Instagram, or Fiverr. Scammers may disappear after receiving a deposit or deliver AI-generated art as if it were original. Genuine artists typically possess sketches, layered files, or work-in-progress images, which AI-generated portfolios lack.
Fake fundraising appeals are also on the rise, with fabricated images of sick children, injured animals, or families in crisis being circulated to solicit donations or gain attention. These images can depict entirely fictional individuals, and their emotional impact often discourages recipients from verifying their authenticity.
While visual inspection is no longer sufficient, some clues can still be observed, such as inconsistent jewelry, unusual lighting, distorted reflections, or odd movements in videos. However, the absence of obvious errors does not guarantee an image's genuineness.
To combat these scams, reverse image searches using tools like Google Lens, TinEye, or Bing Visual Search can help determine an image's origin. However, a lack of search results does not automatically indicate a fake, especially for personal photos or newly published images. A complete absence of history for an image claimed to be circulating for some time or related to a widely reported event warrants suspicion.
Official verification tools are also becoming available. Google's Gemini app, for instance, can detect AI watermarks and provenance data, though it is not infallible. Content Credentials, supported by companies like Adobe and Google, record information about an image's creation or editing. Google's SynthID technology embeds an invisible watermark into AI-generated images, which can survive metadata stripping. Tools like the Gemini app and OpenAI Verify can detect these watermarks.
It is important to note that the absence of a watermark does not confirm an image's authenticity, merely that no watermark was detected. Furthermore, messaging apps often strip embedded credentials when re-encoding images. A valid credential confirms the creation process but not the truthfulness of the content. Third-party SynthID detector sites may also provide estimations rather than direct watermark verification, as direct reading requires specialized technology.
If one suspects they have been a victim of an AI image scam, it is recommended to save screenshots, run the image through verification tools, and contact financial institutions if financial information was shared. Reporting the account to the platform and relevant fraud reporting services is also advised.
