Fraud is a practice that has constantly evolved with the times. New forms of communication
inevitably become new mediums for fraudsters to scam people out of money and valuables.
Like an arms race, as new methods of fraud are developed, so too are new methods of
detection and prevention.
Artificial Intelligence (AI) represents a new paradigm for fraud. While technology enables new
ways to commit and avoid fraud, it ultimately comes down to influencing a person to do
something not in their best interest, and the ultimate weapon for fraudsters remains social
engineering.
Social engineering is the psychological manipulation of people into performing actions or
divulging confidential information, and it is a crucial technique for successful scams. You might
picture a suave con man convincing someone to invest in a phony business opportunity—at its
core, this is social engineering.
Scams often trade off between being convincing and reaching many people. Methods like email
phishing, where fraudsters attempt to collect sensitive information by sending emails that
appear to be from trusted sources, usually work by casting a wide net. By sending phishing
emails in bulk to large numbers of people, fraudsters increase the chance of encountering
someone susceptible to the scam. Some phishing attacks even use emails that aren’t
convincing to most people, ensuring that those who fall for it are more easily manipulated.
Anyone who takes the bait is then put in contact with a real person, who uses social engineering
to acquire what they need from the target.
This is where AI becomes alarming. AI chatbots are now so sophisticated that many people
cannot tell the difference between them and a real person. By leveraging AI, a fraudster no
longer needs to make a trade-off. They can send thousands of convincing phishing emails and
put anyone they trick in contact with an AI chatbot trained to extract money or data. The
advantage is that with AI, fraudsters can cast a wider net and catch more victims while requiring
less human action. This is especially worrisome since some fraudsters target the elderly, who
are more likely to fall victim to scams and have less exposure to AI, making them unlikely to
spot subtle signs of chatbots.
Regardless of how they attract and screen victims, these scams all rely on impersonation. In
impersonation scams, a fraudster assumes the identity of someone with authority, like a bank
representative or IT technician, and exploits the victim’s trust in that position. Typically,
impersonation scams are limited to impersonating authority figures, as it’s difficult to mimic
someone the victim knows personally. However, AI expands the scope of the scam. Recent
advancements now allow AI to clone someone’s voice using just a few minutes of audio and
generate images of them in fabricated scenarios, making it possible to convincingly impersonate
someone the victim actually knows and trusts.
Impersonating a victim’s family member used to be limited to text messages. Fraudsters would
pose as a relative in a time-sensitive situation—maybe a frozen bank account or trouble at the
border while on vacation—but the key factor is an urgent need for money. Scammers often
create a false sense of urgency; when caught up in the story and not thinking critically, victims
are less likely to catch on. With AI, scammers can back up their story with falsified proof, making
it far more convincing. If the scammer pretends to be in an accident on vacation and needs money for hospital bills, they can generate images of their false persona and clone the relative’s
voice to create a message. A family member acting hastily in such a situation can hardly be
blamed, and once the money is sent, there is very little chance of recovering it.
In this digital age, people share so much of their lives and personality on the internet. Mimicking
someone’s face and voice is one thing but combined with the ability of a scammer to study their
mark and build a convincing story, it creates a nearly foolproof deception that can exploit even
the most cautious individuals. The worst part is that many people will be completely convinced
once they think they see the face of their loved one and hear their voice. In those moments,
their natural instincts to protect or help those they care about override any skepticism. The
emotional connection makes it incredibly difficult to question the authenticity of what they’re
experiencing. Scammers exploit this trust, knowing that once the victim feels emotionally
invested, they are more likely to comply with urgent requests—whether it’s sending money,
sharing sensitive information, or even granting access to private accounts. This emotional
manipulation, powered by AI-driven impersonation, leaves victims vulnerable in ways that were
previously unimaginable.
These AI tools represent a massive advancement in the effectiveness of scams. While AI could
technically be leveraged to detect and warn a potential victim of a scam, especially AI-driven
scams, it is inherently more challenging to prevent fraud than it is to perpetrate it. AI systems
can analyze vast amounts of data, identify suspicious patterns, and flag potentially fraudulent
activities with greater accuracy and speed than human intervention. However, fraudsters can
also use AI to enhance their tactics, and they often have the advantage of innovation and
unpredictability. Scammers can constantly adjust their methods, making it easier to bypass
defences, while fraud prevention systems must be reactive, identifying and adapting to new
threats as they arise. Worse yet, the complexity of AI-generated scams, which often involve
deepfake voices, images, and personalized social engineering, makes detection increasingly
difficult. In this constant game of cat-and-mouse, the perpetrator only needs to find one
vulnerability, while defenders must cover every possible angle, making fraud prevention an
inherently more difficult and resource-intensive task.
If you have been the victim of a scam and would like to discuss your legal options, please
Contact our office today to set up a consultation
