Bristol-based expert warns of AI impersonation: Could you identify a fake call from a relative?
Many Britons struggle to identify AI-generated communications
As technology advances, the risk of impersonation scams rises, with many Britons reporting encounters with fraudsters posing as trusted individuals or organisations.
Recent research highlights that although a staggering eight in ten Britons have experienced impersonation attempts, only 14% are very confident in identifying AI-generated or manipulated communications.
Despite these alarming statistics, only 10% of households have agreed on a safe word to verify unexpected requests, leaving many vulnerable to deception by AI.
The Institution of Engineering and Technology (IET) is urging households to adopt private safe words as AI makes it easier to imitate familiar voices and craft convincing messages, images, videos, and documents using personal information gathered online.
Patrick Capaldo from Bristol, Chair of the IET’s AI Technical Network and an AI engineer at Airbus, warns about how AI is revolutionising fraudsters' methods, explaining why familiar voices, numbers, and personal details can no longer serve as reliable identity proof.
"AI has advanced rapidly, enabling fraudsters to mimic a voice with mere seconds of speech," Capaldo explained.
Capaldo emphasised the need for awareness and community vigilance: "We need to normalize sending scam information just so we're all on the same page. Traditionally, targets have been young or elderly people. But now, AI has levelled the playing field—we all can be targets. It's about protecting all of us."
According to IET research, 45% have some confidence in spotting AI-altered communications, yet many traditional indicators of authenticity can now be easily fabricated.
Phone calls remain the most common channel for impersonation attempts, with nearly 29% of those targeted receiving daily suspicious calls. Additionally, people report an average of 26.5 phone impersonation attempts per month, surpassing the rate of suspicious texts (22.8) or emails (21.2).
These cues, once trusted to establish authenticity, are becoming increasingly simple for scammers to replicate using AI. A fraudster might emulate a relative’s voice, mention a genuine event discovered online, and use exposed contact or account details to support an urgent request for money.
Patrick Capaldo stresses the importance of changing the way we verify communications and fostering real-world connections: "Take 20 to 30 seconds to pause and think — reach out through a different channel to verify unexpected requests."
The research reveals a general inconsistency in cyber safety habits, with only 13% of respondents contacting the person or organisation through another channel after an unexpected call, and a mere 13% discussing scams with older family members.
The emotional impact can be significant, with victims experiencing feelings of anger, foolishness, anxiety, and shame. Almost 29% become more cautious in responding to unexpected communications.
Capaldo advocates for awareness: "We need to normalize sharing educational insights on fraud scams within our circles to stay vigilant."
With AI advancing rapidly, Capaldo outlines the need for cautious skepticism: "The digital world offers great opportunities, but we must protect our private offline memories from AI-generated imposters."
Ultimately, the IET encourages everyone to consider their digital footprint: "Knowing what personal information is accessible online can empower individuals and give them confidence in managing their digital presence," Capaldo advises.