How AI voice scams and fake calls work

Criminals use artificial intelligence to scam you

Scam calls are nothing new. Criminals have weaponized the telephone to fraud victims well before the internet was widely available for nearly everyone on the planet. With new technologies, like artificial intelligence, scammers have new tools to target a larger group of victims and make their attacks more convincing — and potentially more lucrative.

More specifically, online criminals can use generative artificial intelligence (AI) to fabricate convincing-sounding calls. Therefore, generative AI and so-called AI voice cloning have made scam calls more difficult to identify as fake. And as AI tools become more accessible and easy to use, the number of AI scams and other such frauds will only grow.

One advantage of generative AI for online criminals is that they can create more convincing scam calls at an unprecedented scale. Until now, scammers have relied on generic attacks that are easy to spot as fake, such as phishing messages with only the recipient’s name changed. With AI, scammers can generate unique scam calls and messages targeted at a certain individual with a much smaller effort.

Generating a person’s voice with AI

With the help of generative AI tools, scammers can pick a snippet of somebody’s voice and create a life-like audio clip of whatever they wish. The AI-generated voice can be eerily reminiscent of the real person, such as your family member. With a sense of urgency mixed in, the victim of an AI scam call may not know to suspect a scam, even if they are familiar with common tricks used by fraudsters.

According to the US Federal Trade Commission (FTC), imposter scams are among the most common, taking the largest slice of the total $8.8 billion victims lost to scammers in 2022. With AI taking over, this number is likely to become much higher. Even the 2022 numbers of consumer losses showed an increase of more than 30% over the previous year.

Fabricating someone’s voice can be done with just a few-second snippet, and there is no shortage of material online. For example, social media is abundant with videos of people talking, making it a treasure trove of material to exploit for scammers. Online criminals can also use voicemail messages or call you directly to get a sample of your voice.

Common AI scams

Voice cloning scams are only one way criminals use artificial intelligence to deceive their victims. In addition to new forms of online fraud, AI can be used to supercharge more traditional scams, such as phishing and identity theft. A few common types of AI scams to look out for include:

• Deepfakes. AI can replicate more than just text and sound. Deepfakes are AI-generated videos and images of real people with their faces pasted on somebody else’s body. Deepfakes are used to fabricate explicit videos of non-consenting individuals, spread disinformation and scam unsuspecting victims.
• Chatbot scams. With AI, online criminals can create chatbots that interact with the user to deceive them. Chatting with these fake people may feel authentic, although in reality, they are trying to steal your identity or scam your money.
• Social engineering. More traditional social engineering attacks aim to get you to do something that compromises your devices to malware or puts your personal information at risk. AI-powered social engineering attacks can come off as more convincing and are easier to pull off on a larger scale.

How to identify fake calls and avoid AI scams

You do not need to have a large bank account to get scammed on the internet. Everyone is a tempting victim of online crime, whether that means hacking, identity theft, spreading malware, or scamming with fake AI calls. Follow this advice to avoid getting swindled with AI scam calls or deceived in some other way:

• Confirm the caller’s identity. Most scams involve impersonation, but AI scam calls make it easier to get deceived. A life-like sound of your loved one combined with a sense of urgency can get you to act carelessly, so double-check that the caller is who they claim to be.
• Reporting scams, including phishing emails, fraudulent SMS messages or AI voice scams, helps others avoid falling for online criminals’ tricks. Scammers often impersonate trusted entities, such as banks, the police or delivery services.
• Use two-factor authentication to protect your accounts. With 2FA turned on, online criminals have a harder time accessing your account, even if they succeed in stealing your pass­word. Even with 2FA, make sure to use strong pass­words.
• Get familiar with AI tools like ChatGPT. Once you are aware of the capabilities of generative AI, you know to be more suspicious of even legitimate-seeming messages and other potential scams.
• Think before you act. You will only face the consequences of AI voice scams and vishing if you act according to the fraudsters’ wishes. Even if the voice sounds like the person you know, verify the caller’s identity by calling their phone number, for instance.
• Do not give AI any personal or confidential information, even knowingly. There have been cases of users’ information being compromised because they told an AI chat tool something confidential.

Although AI technology used for fraud and phone scams gets more nuanced and technologically advanced, so do the tools to combat the malicious use of AI. In addition to being wary of AI scams on the internet, protect your devices against malware with award-winning anti­virus software, such as F‑Secure Total.