Your Career Is One Fake Video Away From Ending: AI Can Now Put Any Bahamian Official Into a Child Abuse Video – And No Court Could Tell It’s Fake

By Michael Elijah Muhammad Clare
10 November 2026

If you hold office in this country, think of the worst thing anyone could ever say about you. Now picture it as a video, with your face, your voice and a child in the frame. You never did it. It never happened. And it is already on a hundred phones.

That is not a claim about any person. No such case has been reported in The Bahamas, and nobody mentioned in this article has been targeted or accused of anything. It is a description of what the technology can now do, taken from police operations, court filings and official figures in Europe, the United States and Asia. Every source is linked at the bottom so you can check it yourself.

It has already started here, in a milder form

On 22 April 2025 the Office of the Prime Minister warned the public that a video appearing to show Prime Minister Philip Davis promoting an investment scheme was not real.

“…fake and generated using artificial intelligence (AI) technology.” — Office of the Prime Minister, 22 April 2025

The matter was referred to the Royal Bahamas Police Force. Days later the Minister of National Security told a press conference that AI video and audio had also been used to impersonate the Governor General and other senior figures in scams.

“These are not harmless tricks.” — Minister of National Security, May 2025

Those fakes were built to steal money. The same cloning of a face and a voice can be aimed at a reputation instead.

In February 2026 the head of the police digital forensics unit told a public forum that high-school students were increasingly being caught in online scams involving child pornography and sex. Police did not say those cases involved AI, and this article does not claim they did.

What the machines can now do

Mainstream products show how far things have moved. OpenAI’s Sora and Google’s Veo produce lifelike video, with matching speech and sound, from a typed description. Those companies build in blocks and watermarks. The criminal market does not.

The Internet Watch Foundation, the UK body that finds and removes child abuse imagery, recorded 13 AI-generated child sexual abuse videos in 2024. In 2025 it recorded 3,443. Nearly two thirds were in the most severe legal category.

The Foundation says much of this material is now indistinguishable from real photography, that sexual video can be generated of anyone with a publicly available image, and that a convincing likeness of a specific child can be built from as few as 20 pictures.

Now count the public photographs of you. Every sitting of Parliament. Every press conference. Every church anniversary.

Five cases you can check for yourself

1. San Francisco, August 2024. The City Attorney sued the owners of 16 of the world’s most-visited websites that turn ordinary photos of women and girls into fake nude images, saying the sites had drawn more than 200 million visits in six months. Read the City Attorney’s announcement.

2. Meta, June 2025. The owner of Facebook and Instagram sued a Hong Kong company behind a family of such apps, saying its adverts kept returning to Meta’s platforms after being removed. Read Meta’s statement.

3. Europol, February 2025. Operation Cumberland, led by Danish police, hit a paid online platform that sold fully AI-generated child abuse imagery to users worldwide. Europol reported 25 arrests across 19 countries and 273 suspects identified. The Bahamas was not among the countries listed. Read Europol’s release.

4. Internet Watch Foundation, 2025 figures. The jump from 13 videos to 3,443, set out above. Read the Foundation’s report.

5. South Korea, 2024. Chat groups trading fake sexual images of women and girls were linked to more than 500 schools and universities. Parliament made it a crime even to possess or watch such material, and police reported 963 people apprehended in a crackdown that ran to March 2025. Read the BBC report.

Britain reached the same conclusion. Measures first announced in February 2025, which the survivors’ charity NAPAC called “a crucial step in safeguarding the next generation”, are now written into the Crime and Policing Act 2026. They include up to five years in prison for AI tools built to produce child abuse images. The FBI has warned since 2023 that criminals turn ordinary photos into explicit fakes to extort victims.

Could a court tell it is fake?

By looking at it, no. A 2024 review of 56 studies found people correctly identify deepfake video about 57 percent of the time. A study of 1,276 people, published in Communications of the ACM in 2025, found people did little better than a coin toss.

Even the specialists struggle. Internet Watch Foundation analysts say they often label material as AI-generated only when the source or the file data gives it away.

Forensic examiners have more than their eyes, and a fake can sometimes be proven fake. But that takes time, money and expertise. No published Bahamian court direction on handling deepfake evidence was found. In England and Wales, judges were formally warned in October 2025.

“Judges should be aware of this new possibility.” — AI Guidance for Judicial Office Holders, England and Wales, October 2025

In the United States, a federal rules committee is still studying a draft rule on deepfake evidence. It has not been adopted. AI has already reached Bahamian court filings: in May 2026 local press reported that an attorney was referred to the Bar’s ethics committee over three AI-generated fake cases.

Then comes the second edge of the blade. Law professors Robert Chesney and Danielle Citron named it the liar’s dividend: once the public knows video can be faked, the guilty can wave away real evidence as fake. The Internet Watch Foundation says offenders are already trying this “deepfake defence”. So justice is hit twice. The innocent can be smeared with a fake, and the guilty can hide behind one.

Risk assessment: the offices in the firing line

What follows is risk, not reported fact. We looked for a published Bahamian case of an official framed this way and found none.

The Prime Minister, Cabinet ministers, Members of Parliament and candidates. A fake released in the last days of a campaign could not be disproved before the vote.

The Commissioner of Police and every officer under that command. One clip, and every arrest that officer ever made is questioned.

The Attorney General and prosecutors. The people who bring the charges could become the accused in the court of public opinion.

Judges and magistrates. A fake on the eve of a ruling invites demands to step aside, and doubt over every past judgment.

No conviction is needed. Under section 16A of the Sexual Offences and Domestic Violence Act, producing child pornography carries life and possessing it carries up to 20 years. An allegation alone can end a career.

That section covers a visual representation of a person who “is, or is depicted as being” under 18. That wording may reach AI-generated material, but no published Bahamian judgment testing it was found.

The May 2026 Speech from the Throne promised cybersecurity legislation, a National Cybersecurity Authority and an Artificial Intelligence Governance Act. As of this writing, no enacted Bahamian offence aimed specifically at deepfakes was found.

So here is the self-interest. The people most exposed are the same people with the power to write, enforce and interpret the law. Pass it for the children. If that is not reason enough, pass it for yourselves.

If it happens to you

• Do not pay. The FBI says paying does not buy silence.
• Keep usernames, links and messages. Never save or forward sexual images of children to “collect proof”. Report instead.
• Report to the Royal Bahamas Police Force on (242) 322-4444 or through its report-a-crime page.
• Report the account to the platform.
• Adults can use StopNCII.org. For images of anyone under 18, use Take It Down.

The video you may never see

WhatsApp and Signal are end-to-end encrypted. Only the people in a chat can read it. That protects ordinary families, and it also protects whoever is passing a lie around.

Europol’s 2026 threat assessment says encrypted messaging apps are widely used by offenders to share illegal material. European police chiefs warned in 2024 that such encryption blinds them to crimes including child abuse, human trafficking, drug smuggling and terrorism. The FBI has warned of violent online networks that trade abuse material and blackmail victims through social media and mobile apps.

So understand this. Right now, a fake video of you could be moving from one private group to the next. You would not see it. Nobody would call. By the time you hear the rumour, the verdict would already be in.

Stop. Question. Verify. Do it before you forward the next clip, because the next face could be yours.

Sources

• Office of the Prime Minister of The Bahamas, press statement on fake AI-generated video, 22 April 2025 — link
• Bahamian press report of the Minister of National Security’s press conference, 2 May 2025 — link
• Bahamian press report of police remarks at a Safer Internet Day forum, 11 February 2026 — link
• Bahamian press report on the planned AI Governance Act and the attorney referral, 21 May 2026 — link
• Office of the Prime Minister, Speech from the Throne, 20 May 2026 — link
• Laws of The Bahamas, Sexual Offences and Domestic Violence Act, section 16A (UNODC copy) — link
• San Francisco City Attorney, lawsuit announcement, 15 August 2024 — link
• Meta, statement on its lawsuit and other action against “nudify” apps, 12 June 2025 — link
• Europol, Operation Cumberland release, 28 February 2025 — link
• Internet Watch Foundation, “Harm without limits” AI report, 2026 — report and 2025 data page
• BBC News, South Korea school deepfake report, 3 September 2024 — link
• CBS News, South Korea law on possessing or watching sexual deepfakes, September 2024 — link
• Asia Economy, Korean National Police Agency crackdown results, April 2025 — link
• NAPAC, “AI Legislation: A Crucial Step in Safeguarding the Next Generation”, February 2025 — link
• UK Government, Crime and Policing Act 2026 child sexual abuse material factsheet — link
• FBI public service announcements, 5 June 2023 and 29 March 2024 — 2023 and 2024
• IBTimes UK, report on the Sora 2 and Veo 3 video models, October 2025 — link
• Diel and others, review of 56 deepfake detection studies, Computers in Human Behavior Reports, 2024 — link
• Cooke and others, “As Good as a Coin Toss”, Communications of the ACM, 2025 — link
• Courts and Tribunals Judiciary, AI Guidance for Judicial Office Holders, October 2025 — link
• US Courts, Advisory Committee on Evidence Rules agenda book, October 2026 — link
• Chesney and Citron, “Deep Fakes: A Looming Challenge for Privacy, Democracy, and National Security”, 2019 — link
• Europol, Internet Organised Crime Threat Assessment 2026 — link
• Eunews, report on the European police chiefs’ declaration on encryption, 22 April 2024 — link
• FBI, public service announcement on violent online networks, 6 March 2025 — link
• Royal Bahamas Police Force, report a crime — link