Eight years of English-language news coverage from 21 countries was analysed to see how the media talks about child sexual exploitation and abuse. Most of the language that child protection experts recommend is rarely used. Some of the language they say causes harm is still common.
In 2016, an international working group published the Luxembourg Guidelines, establishing clear and specific language for talking about child sexual exploitation and abuse. Eight years on, how much of that language has entered mainstream media coverage? The answer is not encouraging. Most recommended language is barely used. Some harmful language is still common.
Of the 88 Luxembourg terminology terms tracked across the analysis, nearly three quarters were either not used at all or appeared so rarely as to have no meaningful presence in coverage. Only 16% of terms appeared frequently. The distribution is stark.
The guidelines exist. The journalism has not followed.
The most frequently used terms fall into two groups. Some are good news: "child sexual abuse," "child sexual exploitation," "grooming," "victim," and "survivor" all appear consistently. But "child pornography" also appears frequently, averaging 41 appearances per year. The Luxembourg Guidelines are explicit: this term should be replaced by "child sexual abuse material" because the word "pornography" implies consent and frames documented abuse as a content category rather than a crime.
Teal bars are recommended terms in good standing. Red is a term the Luxembourg Guidelines say should be replaced. Green bars are terms showing positive growth. "Child pornography" appears at roughly the same annual rate as "survivor."
Some of the most specific and important Luxembourg terminology never appeared once across eight years of coverage from 21 countries. Not once. The most urgent absence is "live online child sexual abuse" , one of the fastest-growing forms of exploitation globally, which has not entered mainstream journalism at all despite coverage of individual cases.
The complete absence of "live online child sexual abuse" from eight years of international coverage is the most urgent finding. Journalists cover individual cases of live-streamed abuse but use generic terms like "live" or "webcam," preventing readers from connecting cases to a recognised pattern.
The analysis is not uniformly discouraging. Six terms showed meaningful and consistent growth from 2016 to 2023. They demonstrate that language change in journalism is possible when professional guidance connects with reporting practice. It is just happening much more slowly than the scale of the issue warrants.
Grew from 1 appearance per year in 2016 to 39 in 2022. Still well below "child pornography" but the direction is clear.
Grew steadily as reporting on digital exploitation increased. Reflects growing journalistic awareness of the online dimension.
Grew from 6 appearances in 2016 to 28 in 2022. A genuine public awareness success. The word is now part of common understanding of how abuse happens.
Growing use alongside "victim" reflects shifting editorial standards. In 2022 and 2023 it reached its highest levels in the dataset, though still at roughly a third of the rate of "victim."
Grew alongside online abuse coverage, with exploitation framing outpacing abuse framing. Both are moving in the right direction.
An emerging term that barely existed in 2016 and was significantly present by 2022. Its rise tracks the growth of the crime itself.
CSAM usage trajectory from 2016 to 2022, showing growth from near-zero to near-parity with the term it is meant to replace. This is what successful terminology adoption looks like.
The most telling number in the dataset is the ratio of "victim" to "survivor." Victim averages 140 appearances per year. Survivor averages 41. Both terms are appropriate in different contexts: victim is right in legal and incident settings; survivor is right in recovery and advocacy contexts. A near four-to-one ratio suggests journalism frames these stories overwhelmingly through the lens of the crime, not the lens of the person. Children are portrayed primarily as things that bad things happened to, rather than as people navigating what happened.
The ratio is improving. "Survivor" reached its highest levels in 2022 and 2023. But at the current rate of change, it will take years before the two terms reach parity in mainstream journalism.
The Luxembourg Guidelines were not developed as an academic exercise. They were developed because language affects outcomes. When a child is described as a "prostitute" rather than a victim of exploitation, it shapes how police, judges, and social workers perceive that child. When "child pornography" is used instead of "child sexual abuse material," it shapes how the public thinks about the severity of what is being distributed online.
Eight years after the guidelines were published, English-language news media has not made the language shift they call for, with a few important exceptions. The terminology gap is not because the guidelines are unknown. They are widely cited in professional child protection circles. The gap exists because journalism, public communication, and professional practice remain largely disconnected.
Methodology. This research was conducted as part of the Childlight Global Data Fellows Programme at the University of Edinburgh. Across 2,024 English-language news articles collected from 21 countries in Africa, Asia, North America, Europe, Australia, and international broadcasters, 88 Luxembourg terminology terms were tracked year by year from 2016 to 2023. All term frequencies are drawn from full article text, not headlines. Findings were presented as a poster at the ISPCAN World Congress. Full methodology documentation available on request.
Mwangaza Watch produces structured briefings on human security across Africa.
All public briefings are open access. mwangazawatch.org