On 14 July 2026, a Cayman Islands resident answered a video call from a man who introduced himself as an immigration officer. He asked her to hold her passport up to the camera. Then he asked for her banking details. She ended the call before answering either question and reported it to the Royal Cayman Islands Police Service. It was the first of what would become four separate public warnings Cayman law enforcement issued about impersonation scams over the following ten weeks.
The tactic did not stand still. By August, callers were switching to Google Meet's audio-only mode, so a victim could never see that no real uniform was on the other end, and some calls displayed the logo of a well known local institution on the caller's screen. Detective Constable Michael McNeir of the Royal Cayman Islands Police Service's Financial Crime Investigation Unit confirmed his team logged at least two of these audio calls in a single day that month. On 19 September, the Cayman Compass reported the police service's fourth public warning about the pattern.
AI can flag a fraudulent video call impersonating police or a bank within seconds. A synthetic-voice classifier scores the caller's audio for artefacts no human ear reliably catches, a caller-verification check cross-references the number or account against a registry of confirmed institutional contacts, and a pattern-matching model compares the script against scams already logged with the Financial Crime Investigation Unit, all before a resident reads out a card number.
None of that is a hypothetical fix for a hypothetical problem. Every capability described above already exists inside fraud-detection products sold to banks in North America and Europe today. What is missing across most of the Caribbean is not the technology. It is the decision to point existing tools at a pattern that is now, by Cayman's own count, four warnings deep.
Four Warnings in Ten Weeks
McNeir's August briefing covered more ground than the video calls alone. Scammers were also targeting Cayman residents shopping for used cars from overseas, asking buyers to wire a 25% deposit over WhatsApp to a contact based in Pakistan for a vehicle that never arrived. In a separate case the same month, a resident who responded to a text message impersonating the Cayman Postal Service, claiming a parcel needed a small delivery fee, had their full card details harvested and used for an unauthorised purchase abroad within hours.
By 28 August, the Royal Cayman Islands Police Service was treating the pattern as a fresh wave, not a continuation of the July case, pairing a renewed Google Meet warning with the parcel delivery text alert. The 19 September warning made four distinct public alerts in ten weeks about what is, underneath the shifting delivery method, the same underlying tactic: a caller or a text borrows the authority of a real institution to extract money or data before the target has time to verify anything.
McNeir put the stakes plainly to reporters in August. "Just remember what Cayman is," he said. "The fifth-largest financial centre in the world." Residents who report a scam quickly have a real chance of getting money back, including a six figure sum his unit recovered in one case. Reported late, that money is usually gone for good.
A Regional Pattern, Not an Island One
Cayman's experience sits inside a wider regional trend. The Eastern Caribbean Central Bank recorded phishing attacks against customers of two commercial banks between January 2023 and March 2024 that cost at least EC$765,000. The Bank of Jamaica logged 443 consumer complaints in 2025, 87 of them fraud related. The Central Bank of Trinidad and Tobago reported that cybersecurity incidents more than doubled between 2023 and 2024.
The exposure is growing because the region's financial habits are moving online faster than the controls built around them. The Bahamas processed B$10.4 billion in retail clearing transactions in 2024. Barbados saw electronic transfers rise more than 700% between 2013 and 2024. Belize recorded a 75.3% jump in instant transfers in 2025 alone, and Guyana moved roughly G$881.7 billion through electronic funds transfers. Each of those figures represents Caribbean residents who now expect a legitimate bank alert, delivery notification, or government email to land on a phone screen. A scammer impersonating any of those senders is counting on exactly that expectation.
Cayman police have also noted the wave is not confined to the Caribbean. Similar Google Meet impersonation scams have surfaced in the United Arab Emirates, Bahrain, Oman, and Finland, part of a pattern large enough that one international operation against related networks froze roughly US$235 million and led to more than 200 arrests. No single island jurisdiction can dismantle infrastructure at that scale on its own. What each one can do is recognise the pattern faster once it arrives locally.
What AI Adds to the Threat
The tools that make this wave harder to stop next year are largely the same tools that could stop it this year. Voice cloning software now needs only a few seconds of someone's recorded speech, pulled from a public interview clip or a voicemail greeting, to generate a convincing imitation. A scammer reading from a generic script today could plausibly call tomorrow in a voice that sounds like a specific named officer or bank manager, because that person's voice is often already sitting in a public news clip or a community meeting recording. Screen-sharing overlays and deepfake video could eventually let a caller display something closer to a genuine uniform, rather than relying on the audio-only workaround Cayman scammers are already using to hide the absence of one.
Small island populations sharpen the risk rather than reduce it. A scammer who scrapes a handful of local news photos and LinkedIn profiles can plausibly reference a victim's employer, neighbourhood, or a mutual acquaintance, details that carry far more weight in a community of a few tens of thousands than in a city of seven million. That same closeness becomes an advantage the moment detection work is automated instead of left entirely to individual judgement under pressure.
Where AI Can Close the Gap
Real-Time Synthetic Voice Detection
Fraud-detection vendors serving banks in North America and Europe already sell models that score a live call for the acoustic signatures synthetic speech leaves behind: gaps in breath timing, unnatural pitch stability, and artefacts around consonants that cloned voices still struggle to reproduce cleanly. Built into a Cayman bank's call-answering system, or offered to residents as a free screening tool tied to the Financial Crime Investigation Unit, that same technology could have flagged McNeir's August Google audio calls as synthetic before a caller ever asked for banking details.
Pattern Matching Against the Unit's Own Case File
The Financial Crime Investigation Unit already holds four documented warnings' worth of scam scripts, numbers, and tactics. Add the Eastern Caribbean Central Bank's, Bank of Jamaica's, and Trinidad and Tobago's data from the same period, and the combined case history is enough to train on. A duty officer working from that model could check a new report against known templates within seconds, rather than relying on memory of the July case while responding to the September one. Every additional warning issued becomes training data rather than just another headline.
Regional Fraud-Signature Sharing
None of the region's fraud data currently feeds into a shared system: not Cayman's video call scripts, not the Eastern Caribbean Central Bank's EC$765,000, not Jamaica's 443 complaints, not Trinidad's doubled incident count. CARICOM IMPACS already coordinates cross-border intelligence on other categories of transnational crime. Extending that coordination to a shared registry of scam phone numbers, WhatsApp accounts, and voice signatures would let a number burning through Cayman this month get flagged automatically the moment it resurfaces against a target in Jamaica or Trinidad the next.
The Core Argument
Cayman's Financial Crime Investigation Unit cannot stop a scammer from placing a Google Meet call, and no filter replaces a resident's own decision to hang up and verify independently. What AI changes is how fast a call gets scored as fake, how quickly one island's warning becomes every island's early alert, and how much of four months of case history sitting in a police file gets put to work before the pattern repeats a fifth time.
What Should Happen Next
Two changes would move the region from repeating warnings to shortening the gap between a new scam tactic and a working defence against it.
First, the Financial Crime Investigation Unit should formalise its four warnings into a structured dataset, the call scripts, the numbers used, the logos displayed, instead of leaving the pattern spread across four separate press releases. That dataset is the raw material any detection model would need, and it already exists inside the unit's own files.
Second, CARICOM member states should agree to route confirmed scam infrastructure, phone numbers, WhatsApp accounts, and spoofed caller identities, through IMPACS close to real time, the way the agency already coordinates on smuggling and other transnational crime. A number reported in George Town on a Tuesday should not get a clean run at a target in Kingston or Bridgetown on the Thursday that follows.
What Is Actually at Stake
None of this closes the gap by itself. A shared registry and a voice classifier still depend on someone building and funding them, and the region's central banks and police services have, so far, published warnings faster than they have built shared tools. McNeir's own case file shows the reporting side already works: residents who called in quickly recovered a six figure sum in at least one instance. Nobody in the region has tried a shared, AI-scored early warning system at scale. Whether it would have caught the July call before three more warnings became necessary is unproven. So is whether the institutional gap that let those four warnings run unconnected would simply swallow a shared tool the same way. That is the question Cayman's next warning should help answer, not this one.
Frequently Asked Questions
What is the Google Meet police impersonation scam affecting the Cayman Islands?
Since 14 July 2026, the Royal Cayman Islands Police Service has issued four public warnings about scammers who contact residents through Google Meet video and audio calls, claiming to be police officers, immigration officials, or bank representatives. Callers ask victims to display identity documents or hand over banking details, and some calls display the logo of a well known local institution on the caller's screen to appear legitimate. Detective Constable Michael McNeir of the Financial Crime Investigation Unit confirmed at least two such calls were reported in a single day in August 2026.
Does this scam only affect people living in the Cayman Islands?
No. Cayman police have noted that similar Google Meet impersonation scams have been reported in the United Arab Emirates, Bahrain, Oman, and Finland, part of a pattern large enough that one international operation against related networks froze roughly US$235 million and led to more than 200 arrests. Across the wider Caribbean, the underlying tactic of impersonating a trusted institution to extract money shows up in separate fraud data from the Eastern Caribbean Central Bank, the Bank of Jamaica, and the Central Bank of Trinidad and Tobago, which suggests the pattern operates across the region and not on one island alone.
How can I verify a video call claiming to be from police or my bank?
End the call without providing any information, then contact the institution directly using a phone number you look up independently, never one given to you during the call or in an accompanying text. The Royal Cayman Islands Police Service, Customs and Border Control, and Workforce Opportunities and Residency Cayman have all stated they will never request passport or banking information over an unsolicited phone or video call. If the caller displayed a logo you recognised, that logo alone is not verification, since scammers can and do display real organisations' branding on Google Meet calls they do not represent.
How much money have Caribbean residents lost to phishing and impersonation scams?
No single consolidated figure covers the whole region, but the pieces that exist are substantial. The Eastern Caribbean Central Bank recorded phishing losses of at least EC$765,000 against customers of two commercial banks between January 2023 and March 2024. The Bank of Jamaica logged 443 consumer complaints in 2025, 87 of them fraud related. The Central Bank of Trinidad and Tobago reported that cybersecurity incidents more than doubled between 2023 and 2024. Cayman's Financial Crime Investigation Unit has recovered a six figure sum for local victims in cases reported quickly, though no public total exists for losses from the 2026 video call wave specifically.
What is the difference between the Google Meet impersonation scam and ordinary phishing texts?
A phishing text asks a victim to click a link and enter details on a fake website, with no live person involved. The Google Meet scam is real time social engineering: a caller adapts their script to what the victim says, expresses urgency or authority in the moment, and can push past hesitation in a way a static text cannot. Cayman police have documented both running in parallel during the same 2026 wave, including a spoofed Cayman Postal Service text that led to a victim's full card details being harvested and used for an unauthorised overseas purchase within hours.
Could AI voice cloning make these scams harder to detect?
Yes. Voice cloning tools now need only a few seconds of someone's recorded speech, pulled from a public interview or a voicemail greeting, to generate a convincing imitation. A scammer currently reading from a generic script could plausibly call using a voice that sounds like a specific named officer or bank manager, since that person's voice is often already public through a news clip or community recording. Small island populations make this sharper, because a caller who references a real neighbourhood, employer, or mutual acquaintance sounds more credible in a community of a few tens of thousands than the same claim would in a large city.
Is there a regional policy or law addressing this kind of fraud in the Caribbean?
Individual regulators issue public warnings, which is the primary tool in use today. The Eastern Caribbean Central Bank and the Central Bank of The Bahamas both run public awareness campaigns, and the Royal Cayman Islands Police Service publishes warnings through its Financial Crime Investigation Unit. CARICOM IMPACS, the Caribbean Community's Implementation Agency for Crime and Security, already coordinates cross-border intelligence on other transnational crime categories, but no shared, real time database currently links scam phone numbers, WhatsApp accounts, or voice signatures reported in one CARICOM member state to what police in another member state are seeing.
Where is AI-powered fraud detection headed for Caribbean banks and police over the next few years?
The building blocks already exist in fraud detection products sold to banks in North America and Europe: synthetic voice scoring, caller verification against confirmed institutional numbers, and pattern matching against known scam scripts. Whether Caribbean institutions adopt them at the pace the fraud is scaling depends less on the technology, which is available now, than on whether police units and central banks fund shared systems instead of publishing separate warnings after each new wave. Cayman's own four warnings in ten weeks show the case history already exists to train such a system today.