An unexpected call sounds like your son, daughter or partner. They seem frightened and need money urgently. AI voice cloning scams can make that voice sound familiar, while a caller’s knowledge of your family makes the story more convincing. Families need a simple way to pause, check and respond.
The speed at which voice imitation has become accessible is striking. However, we should distinguish the growing capability from the number of cases in Ireland. There is no verified Irish trend figure in the sources for this article. The threat is serious without claiming that every alarming call uses AI.
How AI voice cloning scams work
Some criminals copy a person’s voice from audio available online and use it in a false emergency call. Others simply pretend to be a relative without using AI. The US Federal Trade Commission’s family emergency scam advice warns that callers may impersonate a loved one, a police officer, a doctor or a lawyer. They often demand immediate payment and insist that the victim keep the call secret.
These are social engineering attacks. That means the caller manipulates a person’s trust and emotions to obtain money or information. A frightened parent may focus on helping their child, rather than testing the caller’s identity. That response is human and understandable. The criminal is responsible for exploiting it.
The FBI’s guidance on AI enabled financial fraud confirms that criminals can use generated audio to impersonate relatives in a crisis. Its 2025 annual report records more than $5 million in reported US losses from distress scams. That is a US figure, not an estimate for Ireland or a measure of how quickly Irish cases are rising.
Why knowing family details makes a false call credible
A caller might know your child’s name, school, job, sport, holiday plans or the name of your pet. They may know who is travelling and who normally pays for family emergencies. This information can come from public posts, comments, videos, profiles or other exposed records. Familiar details can create false confidence, even when the voice sounds slightly unusual.
An Garda Síochána’s Social Media User Safety Guide explains how online sharing can expose personal information. It advises people to use privacy settings and consider their digital footprint. Therefore, avoid treating a caller’s knowledge of your family as proof that the caller belongs to it.
As AI improves, convincing audio, messages and images may become easier for criminals to combine. For AI voice cloning scams, the safest habit is to verify the request through a separate route. Waiting for an obvious robotic voice or a strange accent is unreliable.
Set a family safety word before an emergency
CODA recommends a family safety word: an obscure, unrelated word or short phrase known only to people who need it. A random example is Stonehenge. Do not use that example as your actual word, because it now appears publicly in this article. Avoid birthdays, pets’ names, school names and anything someone can find on social media.
Agree the word face to face or through a trusted channel. Use it when an unexpected caller claims a relative is in distress. Requests for money, secrecy or immediate action make that check especially important. Make sure children, grandparents and other relevant relatives understand the plan. Never post the word online or share it with a caller who cannot first prove their identity.
If the caller cannot give the word, stop the conversation and verify separately. If they do give it, still call the person back using a number you already trust. A word may be overheard, disclosed or stolen from a compromised message. Change it if that happens. The FBI recommends a secret family word, while the FTC recommends contacting the relative on a known number. Together, these steps help families respond to AI voice cloning scams without relying on voice recognition alone.
What to do during a distress call
- Pause. Do not transfer money, reveal a banking code or follow payment instructions while still on the call.
- Ask for the safety word if doing so is appropriate. Do not offer hints or say the word first.
- End the call and contact your relative independently. Dial the number already saved in your phone. Do not use a number or link supplied by the caller.
- Try another trusted contact if you cannot reach the person. Check their whereabouts without repeating the caller’s demand as fact.
- Treat a credible threat to someone’s immediate safety as an emergency. In Ireland, call 999 or 112 and give the Gardaí the information you have.
A genuine relative can wait while you check. If they are truly in danger, independent contact with them or emergency services helps you respond properly. Do not let demands for secrecy prevent that check.
Reduce the information a caller can use
Review who can see family posts, videos and profile details. Think carefully before sharing travel dates, daily routines, a child’s school or recordings of a family member speaking. Ask relatives to do the same, since one public account can expose information about several people. Privacy settings may reduce the material available for AI voice cloning scams. They cannot erase recordings that others have already copied.
Talk through a practice scenario at home. Include family members who prefer texts or cannot use voice calls, and agree an accessible way to verify a distress message. Schools can share this advice with parents and staff without asking pupils to disclose a family’s private word.
If you sent money or shared information
Contact your bank or payment provider immediately and explain that a scam may have occurred. Ask whether it can stop or recall the transaction, and secure any affected account. Keep the caller’s number, messages, payment details and a timeline of what happened. Do not continue negotiating with the caller or blame the person who responded under pressure.
In Ireland, report suspected fraud to your local Garda station. Garda advice on reporting online scams asks people to bring relevant communications and account details. Garda reporting information confirms that 999 or 112 is for emergencies. A report matters even when no money has yet changed hands.
The CODA View
AI voice cloning scams exploit a basic strength of family life, we want to help when someone we love sounds frightened. Public information can make an impostor sound informed, while artificial audio can make them sound familiar. Better detection tools may help in time, but families need a habit they can use today.
Agree a private family safety word, then practise ending the call and checking through a trusted number. The word is a prompt to slow down, not a guarantee of identity. As the technology changes, that independent check will remain useful. For wider conversations, see CODA’s Resources for Parents and Video Guides for Parents and Children.
Children of the Digital Age provides practical online safety education for families and schools. Last updated: 29 September 2026.
Sources and links in order of appearance
- US Federal Trade Commission, family emergency scams: how voice imitation and urgent requests work, and how to verify separately.
- FBI Internet Crime Complaint Center, AI enabled financial fraud: vocal cloning, secret words and limiting exposed audio.
- FBI Internet Crime Complaint Center, 2025 annual report: US reported distress scam losses and the changing impersonation pattern.
- An Garda Síochána, Social Media User Safety Guide: personal information, privacy and digital footprint advice.
- An Garda Síochána, reporting an online scam: local Garda reporting and useful evidence.
- An Garda Síochána, report a crime: station information and emergency numbers.
Send us a message
Send an enquiry, request information or discuss the needs of your school, family or organisation