AI-assisted scam calls are becoming more common and increasingly difficult to identify. That creates particular risks for people who are less confident with technology, as convincing synthetic voices and video can make familiar-looking fraud seem genuine.
This subject became personal after a family member received a call from an AI bot claiming to work for the UK’s Royal Mail. Research into that incident led to a steady stream of videos showing scam callers being challenged, confused and exposed. The clips are entertaining, but they also reveal how accessible these tools have become.
A simple prompt can derail an AI scam call
Scam-baiting creator Kitboga has shown that basic prompt injection can push some AI scam callers into an endless loop. In one example, the bot repeatedly says “kooga-ooga-amen, Albuquerque, New Mexico” after its instructions are manipulated.
The result is absurd, much like the way some AI systems struggle with apparently simple pronunciations, including the phrase WWE. Behind the comedy is a useful reminder: these systems can sound persuasive while still behaving unpredictably when their instructions are disrupted.
Deepfake video scams raise the stakes
Phone scams are not the only concern. Jim Browning has also examined deepfake video fraud, including a clip where a synthetic-looking man struggles to demonstrate three fingers and repeatedly tries to explain the failure.
The awkward performance may be funny, but deepfake scams can be convincing enough to pressure people into trusting a supposed family member, colleague or authority figure. The technology does not need to be perfect to be effective during a rushed or emotional conversation.
For more technology and gaming video coverage, visit RETROCADE’s arcade videos section.
AI fraud is becoming easier to produce
How widespread are these scams? The AI Incident Database project indicates that they are far from isolated incidents. Simon Mylius, an MIT researcher and collaborator with that project, told The Guardian that the technology is now accessible to almost anyone, with effectively no barrier to entry for producing fake content.
That accessibility also affects scale. In an interview with CNET, Bobby Ford, chief strategy and experience officer at social-engineering defence company Doppel, said threat actors can create deepfakes much faster than before. Faster production means scammers can increase the volume of attacks.
How to help family members spot AI scams
People who follow technology closely may be more likely to question an artificial voice or a video impersonation. That does not mean everyone in a household will recognise the warning signs, particularly when a call appears to come from someone they know.
A practical response is to share examples with relatives and friends who may be vulnerable. Send the clips to a family group chat, discuss them with grandparents and check whether relatives who regularly share AI-generated content on social media understand how these scams work.
It is also worth taking a moment to verify unexpected requests through a separate channel. Hang up and call the person back using a trusted number, ask a private question the impersonator could not easily know, and avoid transferring money or sharing codes under pressure.
For readers exploring hands-on technology and hardware projects, RETROCADE also covers arcade news and industry stories, along with practical guidance in the arcade machine buying guide.
Leave a Reply