Three years ago, we wrote about a chilling new scam: criminals using artificial intelligence (AI) to copy someone’s voice and trick their family into sending money. Back then, it felt like something out of a sci-fi movie. Today, victims have already lost over $5 million in 2025 to “distress” scams alone, where criminals use voice cloning to mimic loved ones in emergency scenarios.

Whether you’re hearing about voice cloning for the first time or you want to understand how much worse it’s gotten, here’s what you need to know: what’s changed since we first warned you, and why the old advice isn’t enough anymore.

What is AI voice cloning?

AI voice cloning is technology that can copy the way a person talks. It captures their tone, accent, and speech patterns and then creates a fake version of their voice that can say anything a scammer wants it to say.

When we first covered this topic in 2023, the technology needed a decent audio sample to work. That’s no longer the case. Today, just three seconds of audio is enough to clone someone’s voice, as confirmed by TrendLife researchers. That three seconds could come from a TikTok video, an Instagram story, a voicemail greeting, or even a voice message in a public group chat.

And here’s the part that should worry everyone: studies show that most people cannot tell the difference between a real voice and a cloned one. That means most of us would fail the test.

The scam doesn’t start with the phone call

In our earlier blogs, we focused on what happens when you pick up the phone and hear a loved one’s voice begging for help. That’s still happening, but we now understand more about how scammers set the whole thing up.

The scam often starts with a hacked social media account. Criminals break into someone’s Facebook, Instagram, or other accounts, often through unsafe third-party apps that ask for “Sign in with Facebook” permissions. Once they’re in, they do two things: steal voice samples from the person’s videos, and use the compromised account to add believability to their lies.

This is what makes modern voice cloning scams so effective. The scammer calls a victim’s family member using a cloned voice. If the family member tries to “do the right thing” and call back through the compromised social media account to verify, the scammer answers using the cloned voice, from the real account. It sounds like mom. The message came from mom’s account. But it’s not mom.

Research shows that one in three people who engage with AI-powered voice calls end up losing money, with average losses of over $18,000 across the six countries surveyed.

How big is this problem?

The numbers tell a stark story. AI-powered voice phishing attacks resulted in $5 million in losses in 2025 alone. Globally, scams of all kinds resulted in an estimated US $442 billion in losses in 2024, and the trend is accelerating. The FBI reported $16.6 billion in cybercrime losses in the US in the same year.

These aren’t just statistics. In one high-profile case in 2024, a finance employee at the engineering firm Arup joined what he thought was a routine video call with his company’s CFO and several colleagues. Everyone on the call, faces and voices alike, were AI-generated deepfakes. By the time anyone realized what had happened, the employee had transferred $25.6 million across 15 transactions.

In 2025, the FBI issued a public alert warning that criminals were using AI-cloned voices to impersonate senior US government officials. Voice cloning is no longer a niche trick. It has gone industrial.

Why the old red flags no longer work

Back in 2023, the advice for spotting scams was simple: watch for bad spelling, strange email addresses, and offers that seem too good to be true. That advice is now dangerously outdated.

Today’s AI-powered scams feature perfect grammar, professional design that looks identical to real websites, and personalized messages based on information scraped from your online presence. Scammers use deepfake voices and video that appear completely authentic. They walk victims through multi-step processes that feel legitimate at every stage. A job offer moves to a WhatsApp chat, which leads to a “processing fee.”

As Lynette Owens, Vice President of Consumer Education & Marketing at Trend Micro, has said: the industry needs to move beyond outdated advice and equip consumers for today’s scams.

The new warning signs to watch for include:

• Impersonation of people or brands you trust, using spoofed phone numbers and real logos
• Requests to move a conversation to WhatsApp, Telegram, or another app
• Voice or video messages that seem authentic but come with an urgent request for money
• Emotional pressure through fear, love, or once-in-a-lifetime opportunity
• Multi-step “journeys” that gradually build trust before asking for payment
• Personalized details that make the message feel like it was meant just for you

Voice cloning is now just one piece of a bigger machine

Perhaps the most alarming development is how voice cloning has become part of a fully automated scam assembly line. With accessible AI tools and simple automation platforms, a single person can now build a polished, high-quality scam operation in just hours, often for as little as $60 per month.

Criminal networks known as “Scam-as-a-Service” operations now package AI voice cloning, deepfake video, fake website generators, and targeted messaging tools into ready-made fraud kits. These services lower the barrier so that even unskilled criminals can run convincing scams at massive scale.

As Stephen Hilt, a Trend Micro Forward Looking Threat Researcher, has observed: criminal networks are evolving into semi-autonomous operations where AI handles persona creation, scriptwriting, and message optimization.

And it doesn’t stop at phone calls. Deepfake technology that once required expensive equipment now runs on ordinary laptops and works in real-time during video calls. Today, the old visual giveaways, such as unnatural eye movements, bad lip-syncing, and odd lighting have largely been solved. You can now have a live video conversation with someone whose face, voice, and identity are entirely fabricated.

There’s also a newer threat on the horizon. Governments across the world are building their own AI systems, trained on national data and designed to speak to citizens in local dialects. These are being rolled out for real services like tax help, welfare applications, and healthcare.

TrendLife researchers warn that criminals can copy these same advantages. Imagine getting a phone call from what sounds like a government AI assistant, speaking your local dialect, referencing a real program you applied for, and asking for your bank details to process a payment. Today’s voice cloning scams succeed even when the caller’s accent is wrong and the story has gaps. When scammers gain access to locally trained voice models and real government procedures (much of which is public information), those last remaining clues that something is off could disappear entirely.

How to protect yourself and your family

The defenses that worked three years ago need a serious upgrade. Here’s what actually works now:

• Set up a family code word. Verbally agree on a secret word or phrase with your closest family members that only you would know. Never write it in messages, emails or social media posts in case your accounts are ever hacked. If someone calls claiming to be your child, parent, or partner in an emergency, ask for the code word. A cloned voice can’t answer a question it was never trained on.
  
  
• Verify through a separate channel. If you get a distressing call or message, don’t call back using the same platform. Hang up and call the person’s actual phone number, the one saved in your contacts, or reach out through a completely different app. If their social media account has been compromised, calling back through that platform will just connect you to the scammer.
  
  
• Lock down your social media. Limit who can see your posts, especially videos where your voice is audible. Be cautious about third-party apps that ask to connect to your social media accounts. These connections can be exploited if apps are compromised or request excessive permissions. Review your app permissions regularly and remove anything you don’t actively use.
  
  
• Pause before you pay. Scammers rely on urgency. They want you panicked and acting fast. Any legitimate situation, whether from a family member, a business, or a government agency, can wait five minutes for you to verify. If someone is pressuring you to pay immediately via cryptocurrency, gift cards, or wire transfer, that pressure itself is the biggest red flag.
  
  
• Report it. If you experience a voice cloning scam or any suspicious contact, report it to the relevant authority in your region such as the FTC (US), Report Fraud (UK), the Canadian Anti-Fraud Centre, or ScamWatch (Australia). Reporting helps authorities track and shut down these operations.
  
  
• Use scam detection tools. AI-powered security tools can now fight fire with fire by scanning calls, messages, and links for signs of fraud before they reach you. Trend Micro ScamCheck, for example, can block scam calls and texts, detect malicious links, and even analyze screenshots of suspicious messages to identify scams.

The Bottom Line

AI voice cloning has moved from a curiosity to a crisis. The technology is cheap, accessible, and getting better every day. Only 38% of consumers say they always check suspicious messages before responding, which means most of us are operating on autopilot in a world that no longer rewards it.

The good news is that the most powerful defenses are also the simplest: a family code word, a five-minute pause, and verification through a separate channel. These low-tech habits can stop even the most sophisticated AI-powered scam in its tracks.

Talk to your family about this today. Not tomorrow. Today. Because the next voice clone a scammer creates could sound exactly like someone you love.