AI voice impersonation presents a significant risk, according to Sam Altman's cautionary stance.
In the digital age, the threat of AI-driven scams is on the rise, with almost a third of people in the UK reporting they have been targeted by an AI voice cloning scam in the past year. One such platform that could potentially be used by scammers is ChatGPT.
To combat this growing issue, financial crime specialist Sarah Lennette, from Starling Bank, offers advice on how to stay vigilant and protect your personal information.
1. Verify Identity Before Acting
When receiving a call or message with a request, especially involving money or sensitive information, it's crucial to double-check by contacting the person directly using previously known contact details rather than replying directly or using caller ID alone. For example, send a text or call back via a trusted number to confirm authenticity.
2. Be Cautious of Unexpected or Urgent Calls
Scammers using AI-cloned voices often create a sense of urgency, keep calls short, or pressure you into acting quickly. If something feels off—such as robotic voice tone, odd pauses, or inconsistent speech patterns—hang up and verify by other means.
3. Limit Publicly Available Voice Data
Minimize sharing voice samples online, like avoiding personalized voicemail greetings or posting videos with clear speech on social media, as these can be used to clone your voice.
4. Avoid Sharing Sensitive Personal Information by Phone
Legitimate institutions generally do not ask for passwords, social security numbers, or bank details over calls. If someone asks for this, do not provide it and verify through official channels.
5. Use Technology Safeguards
Block suspicious numbers and consider using apps or phone settings that automatically detect and filter spam or potentially spoofed calls.
6. Spot Signs of AI-Generated Voice
Notice glitches, unnatural speech, or inconsistent emotions in the voice. AI-generated voices may lack natural flow or have strange tonal shifts.
7. Stay Informed and Report Scams
Report any suspected scams to consumer protection agencies such as Fraud.org, the FBI, or the FTC to help authorities take action and alert others.
By combining vigilance, verification, minimizing voice exposure, and technological protections, you can significantly reduce the risk of falling victim to AI voice cloning scams and protect your personal information.
AI technology can be used to create voice clones and 'deepfake' videos, allowing scammers to commit fraud at a large scale. Lennette warns that anyone asking for money or personal/financial details over the phone should be a red flag. She recommends asking questions that only the person you are speaking to would know the answer to, and that aren't available online.
Lennette also advises against ignoring red flags in phone calls, even if the caller seems to know personal details or claims to be someone you know well. It is best to make the most of shared knowledge with the person you think you are speaking to when identifying potential scams.
To prevent voice cloning, Lennette advises avoiding having a public social media presence and only accepting friends and followers who are known. An estimate from the Global Anti-Scam Alliance suggests that $1.03 trillion was lost to scammers globally in 2024. Financial fraud in the UK has risen, with offenses jumping 46% in the last year. Only a third of people know how to spot an AI cloning scam, according to Starling Bank. Scammers only need three seconds of audio to clone voices, according to Lennette.
Lennette suggests setting up a unique, secret safe phrase with close family and friends as an additional security measure. By following these strategies, you can help protect yourself from the growing threat of AI voice cloning scams.
- In light of the rise in AI-driven scams, it is important to use advanced technology safeguards such as apps or phone settings that automatically detect and filter spam or potentially spoofed calls, a recommendation made by financial crime specialist Sarah Lennette.
- As AI technology can be used to create voice clones and 'deepfake' videos, a warning given by Lennette is that anyone asking for money or personal/financial details over the phone should be a cause for concern, and it's best to verify the authenticity of the caller through alternative means before sharing sensitive information.
