Twice in 2026, people in Trinidad and Tobago watched a video of Finance Minister Davendranath Tancoo recommending a foreign exchange trading platform. The voice sounded right and the face moved right. The minister had never said any of it.
On 10 April 2026, the Trinidad and Tobago Securities and Exchange Commission (TTSEC) warned the public about an AI-generated video on Facebook showing the minister promoting an investment opportunity. On 15 June 2026, the Ministry of Finance issued a second advisory about a new version of the same scheme, naming a supposed trading agent, "Nadia Sanchez", and a platform, "Bullfield Holdings". The ministry called the video "completely fake, unauthorised, and malicious" and asked anyone who received it to report it to the Trinidad and Tobago Police Service, Facebook and Instagram.
The same face was used for the same scam twice in one quarter. Once a fraud network has a convincing synthetic version of a trusted official, it can rerun the campaign at almost no cost, so each response has to be faster than the last.
How a deepfake investment scam works
- Source material. Fraudsters collect public footage of a recognisable person: parliamentary sittings, television interviews, social media clips.
- Generation. An AI video model uses that footage to produce new video of the person saying whatever the script requires, in a close copy of their voice.
- Borrowed trust. A stranger promising large forex returns gets ignored. The country's own Finance Minister saying the same thing gets shared and believed by enough people to make the campaign pay.
- Distribution. Paid social media ads and sharing push the video to as many people as possible before a regulator, bank or platform acts.
Both official warnings came after the videos were already circulating. Verification happened after the harm, and that timing is the weakness every measure below is meant to fix.
The wider fraud picture in Trinidad and Tobago and beyond
At CANTO Connect 2026 in Port of Spain, held from 1 to 3 February, Shiva Bisseesar, managing director of Pinaka Consulting, said business email compromise, in which fraudsters pose as vendors or executives to get staff to approve payments, is "very prevalent" in Trinidad and Tobago. "These guys are so good at trying to get you to bypass your existing workflows," he said.
The country's Financial Intelligence Unit has flagged a rise in romance scams run through dating apps and social media, often over months before any request for money. Financial sextortion is an emerging threat among younger users, and ransomware groups now run like businesses, with separate teams for intrusion, negotiation and laundering.
The Financial Action Task Force, which sets global anti-money laundering standards, published a paper on 24 February 2026 finding that 90 percent of the 156 jurisdictions it assessed identify fraud as a major money laundering risk. It cited a 61 percent rise in cyber-enabled scam cases in Singapore over two years. In the United Kingdom, fraud makes up more than 40 percent of recorded crime, and an estimated 80 percent of it is cyber-enabled.
The Trinidad video was a pre-recorded clip. In Hong Kong in 2024, an employee of a multinational firm joined a video call in which the chief financial officer and other colleagues were all deepfakes, and approved transfers of about US$25 million. Every Caribbean minister, bank chief executive and prominent businessperson has enough public footage online to be impersonated the same way.
Four defences that exist today
1. Impersonation monitoring on social platforms
Monitoring services scan social platforms and ad libraries continuously for a named person's face, name or voice, then run suspect videos through detection models that look for the traces generators leave: odd blinking, lighting that shifts across the face, lip movement slightly out of step with audio, and irregularities in synthetic voice.
Detection is imperfect and generators keep improving, so the value lies in speed rather than certainty. A regulator that learns of a fake within an hour of posting can file takedown requests and publish a warning the same day, instead of days later after members of the public have reported it.
2. Content credentials on official video
The C2PA standard, backed by major technology and media companies and known to users as Content Credentials, attaches a signed, tamper-evident record to genuine video and images showing when and by whom they were produced. If the Ministry of Finance, the Central Bank and TTSEC published every official video with credentials, a journalist, platform or citizen could check any clip against them.
A clip without credentials is not proven fake; most genuine video online carries none. But once official bodies credential everything they release, "no credential" becomes a reason to treat a video of the minister as unverified and to check the ministry's own channels before acting on it.
3. Transaction monitoring at the bank
A scam video still needs a bank account, card or wallet to collect money. Anomaly detection systems watch payments in real time and flag transfers to accounts or platforms linked to reported scams, sudden bursts of transfers, or payments out of character for the customer. This defence works even when the video is never caught, because it targets the money.
4. Liveness checks and call-backs for approvals
The Hong Kong fraud worked because a live video call was treated as proof of identity. Biometric liveness checks look for signs of a real person present in real time rather than a recording or synthetic face. For high-value approvals, a simpler rule also works: confirm by calling back on a number already on file, and require a second approver.
The Core Argument
Trinidad and Tobago cannot stop video generation tools from existing. It can shorten the time between a fake appearing and the public being warned, and it can make the payment channel harder to use. Both are available now.
What Trinidad and Tobago is already doing
Around 10 July 2026, Trinidad and Tobago became the 20th member of LAC4, the Latin America and Caribbean Cyber Competence Centre, a regional cybersecurity body supported by the European Union. Minister of Homeland Security Roger Alexander signed the agreement, and his ministry is the host institution. LAC4 Director Liina Areng said the centre was "pleased to welcome Trinidad and Tobago" and that bringing together experience from across the region "enhances cyber resilience". EU Ambassador Cécile Tassin called the accession "more than a milestone". Membership brings shared threat intelligence, joint training and policy support, which matters because these fraud campaigns rarely stay in one country.
An inter-ministerial committee has also been examining how to update the law for AI-enabled threats, with the Attorney General's Office expected to lead any amendments. In the 2026 cohort of the Youth Electoral Success (YES!) Leadership Programme, young participants in a mock parliament debated criminal penalties for using AI to commit fraud or create harmful deepfakes. It was a training exercise, not a bill, but it shows the issue has reached the next generation of policymakers.
I am less certain that detection will keep pace than the vendors selling it suggest. Each generation of video models removes the flaws the last detectors looked for, which is why I would put more weight on credentials and bank-side controls than on spotting fakes.
What regulators, banks and citizens should do next
- Central banks and securities regulators across the region: contract an impersonation monitoring service covering the governor, finance minister and regulator's own brand, with a target of publishing a public warning within four hours of a confirmed fake.
- Ministries of finance and central banks: publish all official video with Content Credentials from the next budget cycle, and state on their websites that they never endorse private investment platforms.
- Commercial banks: add the names, accounts and web addresses from every regulator scam alert to their payment screening within 24 hours of the alert, and share hits with the Financial Intelligence Unit.
- Company finance teams: adopt a written rule that no payment or change of bank details is approved on a video call, email or voice message alone, only after a call-back to a number already on file.
- CARICOM governments: agree a common credential standard for official government and financial communications, so the same check works on a video from Kingston, Bridgetown or Port of Spain.
Frequently Asked Questions
What happened with the deepfake video of Trinidad and Tobago's Finance Minister?
In April and again in June 2026, AI-generated video of Finance Minister Davendranath Tancoo circulated on Facebook, falsely showing him endorsing a forex trading scheme linked to the names "Nadia Sanchez" and "Bullfield Holdings". TTSEC issued a public alert on 10 April 2026 and the Ministry of Finance an advisory on 15 June 2026. Both confirmed the video was fake and asked the public to report it to the police and the platforms.
How do I check whether an investment firm is allowed to operate in Trinidad and Tobago?
Search TTSEC's published list of registrants on its website before sending any money. A firm offering securities or trading services to the public that is not registered is a red flag, however professional its website looks. TTSEC also publishes investor alerts naming known scams, and you can call it to ask about a firm you cannot find.
I sent money after seeing one of these videos. What should I do now?
Call your bank at once and ask it to stop or recall the transfer; the first hours matter most. Report the scam to the Trinidad and Tobago Police Service and to TTSEC, and keep screenshots, messages, account numbers and receipts. Be wary of anyone who contacts you offering to recover the money for a fee, because recovery scams commonly target people who have already been defrauded.
Is it a crime in Trinidad and Tobago to make or use a deepfake like this?
Using a deepfake to take people's money is already covered by existing fraud and computer misuse law, whatever tool was used. What the country is still considering is whether to create specific offences for making harmful deepfakes, which is the work of the inter-ministerial committee. Until that happens, prosecution depends on proving the fraud, not the fake.
How can a company protect itself from deepfake calls and business email compromise?
Set a written rule that payment approvals and changes to a supplier's bank details need a call-back to a number already on file and a second approver. Train staff that urgency, secrecy and a request to skip the usual process are the warning signs, whatever face or voice is asking. Tell your bank which people are authorised to approve large transfers so it can query anything unusual.
Can ordinary citizens tell whether a video is a deepfake?
Not reliably by watching it, since current fakes pass casual viewing. The dependable checks are the source and the offer: look for the video on the official, verified page of the ministry or bank, and treat any public official endorsing a private investment as a scam by default. Report suspect videos to TTSEC and the platform quickly, because early reports speed up warnings for everyone else.