United Kingdom illegal content risk assessment

This is the full form of KinkEU's assessment of the risk that people in the United Kingdom meet illegal content here, made for the Online Safety Act 2023 and following the regulator's published risk assessment guidance. It says, for each kind of illegal content the guidance names, how likely it is here, how much harm it could do, what we do about it today, and what we plan to do. A shorter summary is published as the document United Kingdom illegal content risk assessment: summary, on the Legal page.

How this assessment was made

The guidance asks for four steps, and this document follows them:

• understand the kinds of illegal content to be assessed: the eighteen kinds of priority illegal content the guidance lists, and other illegal content
• assess the risk of each kind, by how likely it is and how much harm it could do, counting what we already do to prevent it
• decide the measures, put them in place, and keep a record
• report, review and update the assessment

Each kind is given one of four levels: negligible, low, medium or high. The level describes the service as it is on the date of this version, not as we hope it will be. Where something is planned and not yet done, this document says so, and gives no date for it.

Whether the United Kingdom is served

A person who chooses the United Kingdom as their region is told that the service is not provided there. That rests on the person's own choice of region, so somebody in the United Kingdom who chooses another country can still use the service. This assessment is therefore written as if people in the United Kingdom use KinkEU.

What KinkEU is, for this assessment

KinkEU is a service for adults aged 21 or over. Much of what is shared is adult and sexual material, made by the members who publish it. These are the parts that matter most for the risk of illegal content:

• members have profiles with pictures, a name and an age, and can follow and befriend each other
• members can upload pictures and films, comment, and forward what others share
• members can write to each other privately and in groups, and some groups are very large
• members can stream live and make calls. The picture and sound of a large live room are encrypted so that we cannot see them, and nothing from a stream is stored
• members can be found by search once their profile has been reviewed, and on a map and by distance only if they turn that on
• members pay each other in Keys for content, subscriptions, shop items and services, and earned Keys can be withdrawn after a review by staff
• approved providers can list services, including sessions in person
• members can make pictures, films, songs and text with the AI Studio

How we judged each risk

For each kind we looked at the features above against the risk factors the regulator lists for services like ours: an adult service, social media, messaging, groups, live streaming, a marketplace, profiles, location, search and links. We counted the measures that are working today and the limits of each. We looked at how reports are handled and at what the regulator and other adult platforms have published.

We have not yet added the figures from our own reports to this version. When the assessment is next reviewed, those figures are added, and any level they change is changed.

Summary of the levels

High:

• grooming of children
• child sexual abuse material in pictures and films
• intimate image abuse
• extreme pornography
• sexual exploitation of adults

Medium:

• links to child sexual abuse material
• harassment, stalking, threats and abuse
• controlling or coercive behaviour
• human trafficking
• fraud and financial offences
• proceeds of crime
• cyberflashing
• other illegal content, taken together

Low:

• terrorism
• hate
• unlawful immigration
• drugs and psychoactive substances
• firearms, knives and other weapons
• encouraging or assisting suicide and serious self harm
• foreign interference
• animal cruelty

Child sexual exploitation and abuse

Any image of a child is forbidden here, and the minimum age is 21.

Grooming is high. Somebody may join by declaring that they are 21 or over and entering a date of birth, and we cannot yet confirm age in a way the regulator accepts as highly effective. The live face check that opens messaging, posting, streaming and spending shows that a real person is present; it does not estimate age. What lowers the risk today: a profile is in no search until it has been reviewed, nobody is on the map unless they choose to be, a first message to a stranger is a request the other person decides about, profile pictures are checked for faces that may be young, and reports about somebody who may be under 21 go first. The level stays high until a highly effective age check is in place.

Child sexual abuse material in pictures and films is high, as the guidance expects for any adult service that lets people upload. Every member who uploads confirms that everybody shown is 21 or over. Every picture and film in a member's library, every picture posted, and every story picture is checked automatically for a face that may be a child's, and is compared against known abuse material. A library item is never shown to anybody else before it is approved, and a finding of a possible child always goes to a person. Material found to show a child is removed, the account is frozen, the evidence is kept, and the material is reported to the authorities. Private messages, voice and video messages, files sent in a chat, and live streams are not checked automatically, by design, because they are private; they are handled through reports.

Links to child sexual abuse material are medium. Nothing can be posted or sent without an account. Links on members' own pages are reviewed, but links in messages and posts are not yet compared against a list of known abuse addresses.

Intimate image abuse

High. Sharing or making intimate pictures of somebody without their agreement is forbidden. Every uploader confirms that everybody shown agreed. A report that a person shown did not agree is handled first, and the material is taken down while it is checked when the report is credible. A copy of another member's earlier picture is found and refused. In the AI Studio, a real member's face may be used only with that member's own permission, which they can withdraw at any time, and withdrawing it deletes what was made with it that is still stored. Everything the Studio makes is marked as made with AI. We do not yet compare uploads against a shared list of reported intimate images.

Extreme pornography

High, because the service is built around adult and kink material. The content policy forbids the acts the law names: strangulation, choking or suffocation in sexual material, acts likely to cause serious injury, sexual material involving an animal or a dead body, and staged non consent. Strong material must be marked by the member who posts it, a mark added by staff cannot be removed by the member, and adult media stays covered until the viewer chooses to see it. Material is removed on report, and withheld in a country where the law there requires it. We do not yet detect these acts automatically.

Sexual exploitation of adults and human trafficking

Sexual exploitation of adults is high and human trafficking is medium. Arranging either is forbidden. Providers must be approved before they are listed, and a provider's own way of being paid is checked automatically for signs of a scam. Sessions in person can be arranged and paid outside KinkEU, where we cannot see them, which is why the level is high. Reports of pressure, coercion or exploitation go first, and one person may hold only one account, so a closed account cannot simply return.

Harassment, stalking, threats and abuse, and controlling or coercive behaviour

Both are medium. Profiles, private messages, live chat and distance can be misused to find and target somebody. Blocking applies to the account and not to a name, so a new name does not get round it, and one person may hold only one account. Members can mute people and words. Nobody is on the map unless they choose to be, and a place is shown to a city unless the member chooses closer. Some members use features for relationships in which one partner has a say over the other's account; anybody may end such an arrangement at any time, and a report of coercion goes first. A review of those features for misuse is planned.

Fraud, financial offences and proceeds of crime

Both are medium. Keys stay inside KinkEU and cannot be given directly from one member to another. A withdrawal of earned Keys is reviewed by staff, has daily, weekly and monthly limits, can go only to one address the member saved and confirmed with a code and that has waited before use, and needs one member of staff to approve it and a different one to send it. One account for each person and the live face check make fake profiles harder. We do not yet screen withdrawal addresses automatically against lists of addresses linked to sanctions or crime; that is planned.

Cyberflashing

Medium. Sexual pictures between adults who want them are what the service is for, which makes an unwanted one more likely. A picture or film in a first message from somebody the receiver has not accepted is shown only as a tiny blurred preview until the receiver chooses to see it. Adult media marked by the sender is covered until the viewer opens it. Blocking, muting and reporting apply. An unmarked sexual picture from somebody already accepted is shown as sent, and covering it by default is planned.

The kinds judged low

Terrorism, hate, unlawful immigration, drugs and psychoactive substances, firearms, knives and other weapons, encouraging or assisting suicide and serious self harm, foreign interference and animal cruelty are judged low. Each has risk factors here, such as groups, private messages and live streams, but none has a connection with what the service is for. Each is forbidden by our rules and removed on report, and a report of a threat to somebody's life goes first and is passed to the authorities. Sexual material involving an animal is forbidden and is handled as extreme pornography.

Other illegal content

Medium, taken together. This includes copyright infringement, such as paid material copied and shared without the maker's agreement, and content that is lawful in one country and unlawful in another. Copyright complaints have their own process, and content found unlawful in one country only is withheld there.

Measures across every kind

• a report press on profiles, messages, pictures, films, posts, comments, events, streams, live chat, groups, bookings and reviews, and a contact form for everybody without an account
• the most serious reports handled first: anything about somebody who may be under 21, a person shown without agreement, coercion, and a threat to life
• an automatic review of every profile and every gallery picture against our written rules, which approves, refuses with a reason the member can read, or sends to a person; a possible child always goes to a person
• removal of illegal content once we know of it, a statement of the decision, the rule it rests on and how to appeal, and an appeal decided by a person
• terms and policies that say what is forbidden, in plain words, in twenty one languages

Measures planned

These are planned. No date is promised for any of them.

• an age check the regulator accepts as highly effective, before adult material or private messages for people in the United Kingdom, without asking for an identity document
• holding every upload the child safety check flags automatically, before a person looks
• comparing uploads against recognised lists of known child sexual abuse material and of reported intimate images
• comparing links against a recognised list of child sexual abuse addresses
• checking every face in a picture given to the AI Studio, as for faces stored with us
• screening withdrawal addresses against lists of addresses linked to sanctions or crime
• covering unmarked sexual pictures in messages by default
• signals to staff for material that may show the acts the extreme pornography law names
• saying in each statement of a decision whether automated tools took part
• a review of the relationship features for coercion
• a written record of the moderation team's size and training, and a regular report on how well each measure works

Who is responsible, and reviews

The service's responsible person holds this assessment, approves it, and reports its findings and the progress of the planned measures. It is reviewed at least once every twelve months. It is reviewed again, or a new assessment is made before the change, whenever:

• the regulator changes its risk profiles or guidance in a way that concerns this service
• we plan a significant change, such as a new way to communicate, a change to search, recommendations, maps or distance, a change to how age is checked or how pictures are checked, a new kind of paid content or payout, or a change to the AI Studio's rules about faces
• reports of any kind rise, a serious incident happens, or the regulator or the police ask
• the number of people using the service in the United Kingdom grows past a level at which different measures apply

Children

The service is for adults aged 21 or over, and is not meant to be used by children. How age is asked and checked, and how children are kept from adult material, is set out in the document on age eligibility and age assurance.

Changes

This assessment carries a version and a date. When it is reviewed, a new version is published, together with a new summary.