Blocking Is Not Enough: Jamaica's AI and Gender-Based Violence Report

UNESCO's June 2026 report addresses harms that move through exactly this kind of ordinary screen: deepfakes, digital stalking, and non-consensual images. Photo via Unsplash.

Jamaica does not have a law built for what a deepfake actually does. UNESCO's Office for the Caribbean and Jamaica's Office of the Prime Minister and Bureau of Gender Affairs spent 29 and 30 June 2026 launching a report that says so plainly. "Blocking is NOT Enough" addresses AI-enabled gender-based violence, deepfakes, non-consensual image generation, digital stalking, and algorithmic discrimination, and its own title is the argument: telling a woman to block an account after a fabricated image of her is already circulating treats the symptom, not the harm.

The gap between that report and Jamaica's existing law is specific and documented. Attorney and data privacy professional O'Neil Corinaldie has said the Cybercrimes Act and Obscene Publications Act can sometimes apply if altered content is used to harass or distress someone, but neither statute was written with generative AI in mind, and most of what they offer a victim is a civil claim, such as defamation, rather than a criminal charge aimed at the deepfake itself. Corinaldie's own estimate for when Jamaica might have a fully structured AI legal framework: around 2035. That is not a rounding error. It is close to a decade after the report that just described the problem.

What Actually Launched on 29 and 30 June

The workshop brought together government officials, civil society groups, academics, youth representatives, and digital governance specialists in Jamaica, under a European Union-funded project supporting member states in implementing UNESCO's Recommendation on the Ethics of AI. UNESCO's regional director and representative, Eric Falt, framed the stakes without hedging: "This is not only an online safety issue, but a human rights issue, a gender equality issue, a digital governance issue." Hon. Olivia "Babsy" Grange, Jamaica's Minister of Culture, Gender, Entertainment and Sport, brought the framing back to the people behind the statistics: "Behind every data point is a real person, a daughter, a sister, a colleague, a neighbour, who deserves to be safe."

Those two quotes describe the same event from two directions. Falt is naming this as a governance problem that belongs in policy and law, not just awareness campaigns. Grange is naming it as a problem that happens to specific, known people in specific Jamaican communities, not an abstract global trend. The report's job was to hold both of those true at once.

The Scale of the Problem, in Numbers

Deepfake tracking research that UNESCO and UN Women both draw on puts the pornographic share of deepfake videos online at around 98%, with roughly 99% of that content targeting women. The overall volume of deepfake video has grown by something like 550% since 2019, a curve driven almost entirely by how much cheaper and faster the underlying AI tools have become, not by any change in who wants to misuse them. Separate global survey data, cited in the same body of research UNESCO draws from, finds that 38% of women report personal experience of online violence, and 85% of women who use the internet say they have witnessed digital violence directed at someone else.

None of those figures are Jamaica-specific measurements, and it would be dishonest to present them as if they were. What they establish is the shape of the problem Jamaica is confronting: a technology that is overwhelmingly weaponised against women specifically, growing fast, and landing in a country where the underlying rates of gender-based violence were already high before AI entered the picture. Regional data on intimate partner violence across the Caribbean sits somewhere between 39% and 55% of women reporting personal experience, depending on the survey and country. AI does not create that baseline. It gives an existing problem a faster, cheaper, and more anonymous delivery mechanism.

Where Jamaican Law Actually Stops

Jamaica is not starting from zero. The Cybercrimes Act addresses the non-consensual sharing of intimate content, and it does so using the broader concept of publishing rather than the narrower idea of sending an image to one other person, which closes an obvious loophole. It also treats anyone under 18 as incapable of giving consent to have their image shared at all, regardless of how the image was obtained. Those are real protections, and they predate most of the current wave of AI concern.

What they do not do is address content that was never a real photograph in the first place. Section 9 of the Cybercrimes Act and the Obscene Publications Act can sometimes be stretched to cover a fabricated image used to harass or distress someone, but Corinaldie is blunt about the limits: "Our laws don't adequately capture the use of deep fakes or similar scenarios," and, more broadly, "we still have a lot of catching up to do." In practice, a woman whose face is convincingly placed into fabricated content right now has a defamation claim to pursue in civil court, which requires money, time, and a named defendant she can actually serve, rather than a straightforward criminal complaint against the act of creating and distributing the material itself.

Protection built after the fact, a civil claim, a takedown request, is what Jamaica's current law mostly offers. The report argues for protection built in from the start. Photo via Unsplash.

Why "Just Block It" Was Never a Real Strategy

The workshop's own framing rejected the advice most women already hear by default: block the account, delete the post, stay off the platform for a while. Those are individual coping tactics, and they ask the person being targeted to absorb the entire cost of managing someone else's abuse. The report organises its response around four areas instead: prevention, accountability, education, and protection, aimed at platforms, institutions, and law rather than at the person being harmed.

People working the front line of this problem describe the gap in plainer terms than a policy document does. As one gender-desk support worker in Kingston put it during a community outreach session tied to the report's launch, "Dem tell har fi just block him, but block cyaan stop di picture from spread once it out deh. We haffi stop it before it reach har phone, not after." That is the entire argument of "Blocking is NOT Enough" in one sentence: by the time blocking is the available option, the harm has usually already happened.

What AI Governance Could Actually Add

A report is a starting document, not an enforcement mechanism, and the honest question is what closes the distance between the two. Some of that work is technical rather than legislative. Detection tools that flag AI-generated or manipulated media before wide distribution, platform-level policies that treat non-consensual synthetic imagery as a removable violation rather than a grey area, and content provenance standards that make fabricated media easier to identify at the point of upload are all available today, not theoretical future technology.

Some of it is governance work in the narrower sense. The Caribbean AI Risk Management Council has built assessment frameworks aimed at exactly this kind of accountability gap, where an institution has adopted or is exposed to AI systems without the controls to govern what those systems can be used to do to a person. The Caribbean AI Association runs training and awareness work across the region that complements what Jamaica's Bureau of Gender Affairs is doing domestically. And AI Jamaica, StarApple AI Jamaica's sister site, has covered the parallel workforce and platform-safety questions this same technology raises in other sectors. None of that replaces legislation. All of it is the kind of groundwork that determines whether a future Jamaican deepfake law actually works when it eventually passes, rather than sitting unused the way older, narrower cybercrime statutes sometimes do.

What a 2035 Timeline Actually Costs

It is worth being honest about what a nine-year gap means in practice, rather than treating the 2035 estimate as an abstract planning detail. The AI tools capable of producing convincing fabricated images have gotten dramatically better and cheaper in the two years since this problem became a mainstream policy concern. There is no reason to expect that curve to flatten while Jamaica's legislative process works through its ordinary pace. A law finalised in 2035, built around the AI capability of that year rather than this one, risks being outdated before it takes effect, in the same way early cybercrime law was written for a internet that looked nothing like the one that exists now. That is not a reason to skip the legislative process. It is a reason the technical and institutional groundwork, the detection tools, the platform accountability, the governance frameworks, needs to start now rather than wait for the law to arrive first.

Bottom Line

UNESCO and Jamaica's government spent two days in June naming a problem clearly: AI is being used to harm women and girls in ways the country's current law does not squarely address, and the honest timeline for a fix runs close to a decade. Falt called it a governance issue. Grange called it a problem that happens to real people. Both are right, and neither claim does anything on its own until the platforms, the institutions, and eventually the law catch up to what the report already said out loud.

Frequently Asked Questions

The systems capable of producing convincing fabricated images keep improving faster than the law that would govern their misuse. Photo via Unsplash.