Scam scan
The scam scan looks at an image for the telltale signs of a scam. A common use is checking an online sales listing. For example, a screenshot of an item for sale on Facebook Marketplace can be scanned to check for signs that the listing is not what it claims to be.
To run a scam scan:
- Tap Upload image.
- Choose the image to check.
- The scanner analyzes it and returns a result.
Depending on the network upload speed and the size of the image, the analysis can take a few seconds to come back.
If the result does not seem right, feedback can be given using the prompt "Did the AI get this verdict right?" with a Yes or No answer. This feedback helps improve the accuracy of future scans.
Deepfake scan
The deepfake scan checks an image for deepfake content and for signs of manipulation that could be used in a scam. This is useful when an image looks convincing but something about it feels off.
For example, if an image arrives over WhatsApp that appears to be fake, it can be saved to the device and then run through the deepfake scanner to check whether it has been generated or altered.
The steps are the same as the scam scan: tap Upload image, choose the image, and wait a few seconds for the result.
A note on results
Both scanners use AI models to analyze images, and like any AI, they can make mistakes. A result is a strong signal, not a guarantee. It is best treated as one part of a wider judgment rather than the final word.
The oldest rule still applies: if something looks too good to be true, it probably is. When a deal, a message, or an image raises doubt, it is worth pausing before sending money, sharing details, or clicking a link.