Professionals can mistake the consequences of exploitation for evidence that exploitation is not happening. When a child does not behave like the victim we expect to see, professional curiosity must increase, not compassion decrease.
A 15-year-old girl repeatedly goes missing. When she returns, she has a phone she cannot explain. She refuses to identify the adults she has been with. She becomes angry when questioned and continues contacting somebody professionals believe is harming her. The language around her begins to change. She is described as defiant. Risk-taking. Secretive. Unwilling to engage. Some records suggest that she is making poor choices or placing herself at risk.
Those descriptions are not necessarily invented. They may accurately record what professionals can see. The problem begins when visible behaviour is treated as a complete explanation.
Another question is possible: what level of control would make returning feel safer than leaving?
That question does not assume that every missing episode, difficult relationship or offence is proof of exploitation. It does something more disciplined. It keeps the possibility of coercion open long enough for professionals to test it properly.
Behaviour is not the same as freedomChildren affected by exploitation do make decisions. They may choose when to answer a phone, what to tell a professional or whether to return to a location. But the existence of a decision does not tell us whether the child was free to make it.
A useful safeguarding assessment asks not only, ‘Did the child do this?’ It also asks, ‘What shaped the range of options available to them?’
Apparent choices may be constrained by threats, emotional dependency, debt, sexual-image blackmail, fear of retaliation, loyalty, shame, concern for siblings or friends, or the belief that nobody else can keep them safe. The Home Office guidance on county lines describes the use of violence, psychological control, blackmail, online entrapment and manufactured debt. It also recognises that child criminal exploitation may be present even where the activity appears consensual.
Exploitation rarely begins with its most frightening demand. It can begin with attention, money, status, protection, affection or belonging. What first feels like an opportunity can become a debt. What feels like a relationship can become surveillance. What looks from the outside like loyalty may, from the child’s position, be a survival strategy.
The Centre of expertise on child sexual abuse makes the same point in relation to sexual exploitation: apparent cooperation cannot be taken as consent where grooming, coercion and control are operating. Online and offline harm can overlap, including threats to share sexual or nude images.
We should therefore be careful with a simple question such as, ‘Why did they go back?’ Returning may not mean the child wants the harm. It may mean the child fears what will happen to them or somebody else if they refuse. It may mean the exploiter has become their main source of identity, protection or emotional connection. It may mean that leaving feels more dangerous than staying.
The Perfect Victim Test
Safeguarding systems can unconsciously operate a perfect victim test. The child most readily recognised as harmed is often the child who discloses early, gives a clear and consistent account, accepts professional language, cooperates with the police, cuts contact with the perpetrator and shows distress in a form adults recognise.
Real children may do none of those things.
A child may minimise what has happened. They may protect the person harming them. They may withdraw one part of their account, disclose another part later or communicate through anger, avoidance, running away or self-harm rather than words. Research summarised by the CSA Centre describes disclosure as a process, not a single event. Verbal disclosure during childhood is often low, some accounts are partial, and children may show through their behaviour what they cannot yet say directly. If our threshold for belief is calm cooperation and immediate disengagement, we will miss many of the children whose circumstances are most controlled.
The language used in professional records then becomes critical. Phrases such as ‘putting herself at risk’, ‘involved in’, ‘choosing this lifestyle’ or ‘refusing to engage’ can move attention away from the adult creating the harm. The Children’s Society’s language guidance warns that victim-blaming language can reinforce shame, create barriers to disclosure and alter how other professionals understand and respond to the child.
Let’s tell the truth: sometimes the child’s presentation is difficult. They may be hostile, unreliable, frightening, harmful to others or exhausted by professionals. Compassion does not require us to pretend otherwise. It requires us to ask what the behaviour is communicating and who has gained from it.
When the child is also causing harmThe hardest cases are often those in which a child is both harmed and harming others. A criminally exploited child may carry drugs, hold a weapon, assault somebody, recruit a friend or introduce another child to an exploitative network. A sexually exploited child may facilitate access to peers or appear to defend an abusive group.
Recognising exploitation does not erase the harm caused to other children. Nor should safeguarding be confused with the absence of accountability. The task is to avoid a false choice between seeing a victim and seeing an offender.
A child can require protection, boundaries, intervention and accountability at the same time. What changes is the quality of our response. We examine who directed the behaviour, who benefited, what coercion or dependency existed, and what would reduce the likelihood of further harm. Punishing the visible child while leaving the controlling adult and the exploitative context untouched is not accountability. It is an incomplete response.
Every agency sees a fragment. Exploitation often becomes visible in fragments.
School
The fragment they may see: Absence, exclusion, falling attainment, new peers or sudden changes in presentation
Police
The fragment they may see: Missing episodes, possession, offending, repeated locations or contact with known adults
Health
The fragment they may see: Injuries, anxiety, self-harm, sexual-health concerns, substance use or repeated emergency attendance
Social care
The fragment they may see: Conflict at home, placement instability, family distress or concerns that do not fit an intra-familial model
Youth workers
The fragment they may see: Fear, new associations, unexplained money, changes in language, status or behaviour
Family
The fragment they may see: Small changes over time, unexplained routines, names, calls, gifts, threats or a child they no longer recognise
The failure happens when each fragment is treated as a separate problem. Absence becomes an education issue. Possession becomes a policing issue. Family conflict becomes a parenting issue. Injury becomes a health episode. Nobody asks whether the same person, place, phone, debt or threat connects them.
This is why multi-agency working is not a meeting held after concern has escalated. It is the active construction of a shared picture. The 2024 joint inspection of multi-agency responses to serious youth violence found that safeguarding needs, exploitation, trauma and wider needs were not always recognised. It also found that stronger partnerships collated information across agencies and responded rapidly, while ineffective information-sharing left disruption reactive rather than preventative.
Families belong in that shared picture. They may not hold every answer, and home is not safe for every child. But where parents, carers or siblings are safe, their observations should not be reduced to anxiety, poor boundaries or an inability to manage adolescence. Families often see the pattern between professional contacts. They notice the sequence, not just the episode.
The current London Assembly Police and Crime Committee investigation into grooming gangs in London is examining prevention, investigation, victim support and partnership working. Those questions cannot be answered by counting incidents alone. They require scrutiny of how signs were interpreted, whose account was believed, what agencies knew and whether anybody joined the fragments together.
Five questions to ask before calling it a choice1. Who benefits from what the child is doing?
Follow the benefit. Who receives the money, status, sexual access, information, transport, premises or new recruits? A child may be the most visible person in the chain while holding the least power.
2. What might happen if the child refuses?
Do not limit this to an explicit threat. Consider violence, humiliation, exposure of images, loss of protection, withdrawal of affection, debt, retaliation against family members or fear of police and services.
3. What threat, reward, debt or dependency is operating?
Map both pressure and attachment. Exploitation can be maintained by fear and by the child’s need for belonging. The reward does not cancel the coercion. It may be one of its mechanisms.
4. What information is held by other agencies or the family?
Ask what changes when the missing episode, school absence, injury, new phone, location, peer group and offence are placed on the same timeline. Identify who is responsible for building and updating that picture.
5. What would need to change for the child to feel safe enough to speak honestly?
Disclosure is affected by safety, trust and what the child believes will happen next. Consider who should speak with them, where, at what pace, and what practical protection can be offered before asking for more detail.
What a better professional response looks like
Greater curiosity must lead to different action, not simply more questions for the child.
A stronger response keeps a consistent trusted professional involved, avoids closing support because engagement is uneven, and builds a safety plan around the places and people connected to harm. It records facts clearly while separating observation from interpretation. It asks what can be disrupted around the adult, network or location, rather than placing the entire burden of safety on the child.
It also avoids repeated interrogation. The child should not have to produce a perfect disclosure for each new professional in order to remain eligible for belief. Information should move safely between agencies so that the child does not carry the system’s communication burden.
And it pays attention to language. Compare ‘she keeps choosing to return to him’ with ‘she has returned to contact with an adult about whom there are concerns regarding coercion and control’. The second formulation does not manufacture certainty. It keeps the safeguarding hypothesis visible and directs attention towards the adult’s behaviour.
Professional curiosity is not about removing all agency from children. It is about refusing to award adult freedom to a child whose choices may have been narrowed by fear, dependency and control.
The question we should leave withThe child who refuses help may be the child who has learned that help cannot protect them. The child who returns may be responding to a threat nobody has yet uncovered. The child who protects an adult may be protecting a sibling, a friend, their reputation or the only relationship they believe they have.
None of this means professionals should assume exploitation whenever a child takes a risk or commits an offence. It means we should be as curious about control as we are about conduct, and as interested in the adult creating the conditions as we are in the child reacting to them.
The question is not simply why the child keeps returning. The question is what the person exploiting them has made them believe will happen if they do not.
Evidence and further reading
Home Office: Criminal exploitation of children and vulnerable adults, county lines. Defines child criminal exploitation and describes coercion, blackmail, online entrapment and debt bondage.
Centre of expertise on child sexual abuse: Child sexual exploitation. Summarises research on control, coercion, apparent cooperation and the overlap between online and offline exploitation.
Centre of expertise on child sexual abuse: Identifying and responding to disclosures. Explains disclosure as a process that may be partial, delayed, behavioural or non-verbal.
The Children’s Society: Appropriate Language, Child Exploitation. Shows how victim-blaming language can affect professional thinking, disclosure and support.
Joint inspectorates: Multi-agency responses to serious youth violence. Reports on identification, information-sharing, trauma, exploitation and joined-up safeguarding practice.
Department for Education: Working Together to Safeguard Children 2026. Current statutory guidance for multi-agency work to help, support and protect children in England.
London Assembly: Grooming gangs in London investigation. Sets out the committee’s current focus on prevention, investigation, victim support and partnership working.
A 15-year-old girl repeatedly goes missing. When she returns, she has a phone she cannot explain. She refuses to identify the adults she has been with. She becomes angry when questioned and continues contacting somebody professionals believe is harming her. The language around her begins to change. She is described as defiant. Risk-taking. Secretive. Unwilling to engage. Some records suggest that she is making poor choices or placing herself at risk.
Those descriptions are not necessarily invented. They may accurately record what professionals can see. The problem begins when visible behaviour is treated as a complete explanation.
Another question is possible: what level of control would make returning feel safer than leaving?
That question does not assume that every missing episode, difficult relationship or offence is proof of exploitation. It does something more disciplined. It keeps the possibility of coercion open long enough for professionals to test it properly.
Behaviour is not the same as freedomChildren affected by exploitation do make decisions. They may choose when to answer a phone, what to tell a professional or whether to return to a location. But the existence of a decision does not tell us whether the child was free to make it.
A useful safeguarding assessment asks not only, ‘Did the child do this?’ It also asks, ‘What shaped the range of options available to them?’
Apparent choices may be constrained by threats, emotional dependency, debt, sexual-image blackmail, fear of retaliation, loyalty, shame, concern for siblings or friends, or the belief that nobody else can keep them safe. The Home Office guidance on county lines describes the use of violence, psychological control, blackmail, online entrapment and manufactured debt. It also recognises that child criminal exploitation may be present even where the activity appears consensual.
Exploitation rarely begins with its most frightening demand. It can begin with attention, money, status, protection, affection or belonging. What first feels like an opportunity can become a debt. What feels like a relationship can become surveillance. What looks from the outside like loyalty may, from the child’s position, be a survival strategy.
The Centre of expertise on child sexual abuse makes the same point in relation to sexual exploitation: apparent cooperation cannot be taken as consent where grooming, coercion and control are operating. Online and offline harm can overlap, including threats to share sexual or nude images.
We should therefore be careful with a simple question such as, ‘Why did they go back?’ Returning may not mean the child wants the harm. It may mean the child fears what will happen to them or somebody else if they refuse. It may mean the exploiter has become their main source of identity, protection or emotional connection. It may mean that leaving feels more dangerous than staying.
The Perfect Victim Test
Safeguarding systems can unconsciously operate a perfect victim test. The child most readily recognised as harmed is often the child who discloses early, gives a clear and consistent account, accepts professional language, cooperates with the police, cuts contact with the perpetrator and shows distress in a form adults recognise.
Real children may do none of those things.
A child may minimise what has happened. They may protect the person harming them. They may withdraw one part of their account, disclose another part later or communicate through anger, avoidance, running away or self-harm rather than words. Research summarised by the CSA Centre describes disclosure as a process, not a single event. Verbal disclosure during childhood is often low, some accounts are partial, and children may show through their behaviour what they cannot yet say directly. If our threshold for belief is calm cooperation and immediate disengagement, we will miss many of the children whose circumstances are most controlled.
The language used in professional records then becomes critical. Phrases such as ‘putting herself at risk’, ‘involved in’, ‘choosing this lifestyle’ or ‘refusing to engage’ can move attention away from the adult creating the harm. The Children’s Society’s language guidance warns that victim-blaming language can reinforce shame, create barriers to disclosure and alter how other professionals understand and respond to the child.
Let’s tell the truth: sometimes the child’s presentation is difficult. They may be hostile, unreliable, frightening, harmful to others or exhausted by professionals. Compassion does not require us to pretend otherwise. It requires us to ask what the behaviour is communicating and who has gained from it.
When the child is also causing harmThe hardest cases are often those in which a child is both harmed and harming others. A criminally exploited child may carry drugs, hold a weapon, assault somebody, recruit a friend or introduce another child to an exploitative network. A sexually exploited child may facilitate access to peers or appear to defend an abusive group.
Recognising exploitation does not erase the harm caused to other children. Nor should safeguarding be confused with the absence of accountability. The task is to avoid a false choice between seeing a victim and seeing an offender.
A child can require protection, boundaries, intervention and accountability at the same time. What changes is the quality of our response. We examine who directed the behaviour, who benefited, what coercion or dependency existed, and what would reduce the likelihood of further harm. Punishing the visible child while leaving the controlling adult and the exploitative context untouched is not accountability. It is an incomplete response.
Every agency sees a fragment. Exploitation often becomes visible in fragments.
School
The fragment they may see: Absence, exclusion, falling attainment, new peers or sudden changes in presentation
Police
The fragment they may see: Missing episodes, possession, offending, repeated locations or contact with known adults
Health
The fragment they may see: Injuries, anxiety, self-harm, sexual-health concerns, substance use or repeated emergency attendance
Social care
The fragment they may see: Conflict at home, placement instability, family distress or concerns that do not fit an intra-familial model
Youth workers
The fragment they may see: Fear, new associations, unexplained money, changes in language, status or behaviour
Family
The fragment they may see: Small changes over time, unexplained routines, names, calls, gifts, threats or a child they no longer recognise
The failure happens when each fragment is treated as a separate problem. Absence becomes an education issue. Possession becomes a policing issue. Family conflict becomes a parenting issue. Injury becomes a health episode. Nobody asks whether the same person, place, phone, debt or threat connects them.
This is why multi-agency working is not a meeting held after concern has escalated. It is the active construction of a shared picture. The 2024 joint inspection of multi-agency responses to serious youth violence found that safeguarding needs, exploitation, trauma and wider needs were not always recognised. It also found that stronger partnerships collated information across agencies and responded rapidly, while ineffective information-sharing left disruption reactive rather than preventative.
Families belong in that shared picture. They may not hold every answer, and home is not safe for every child. But where parents, carers or siblings are safe, their observations should not be reduced to anxiety, poor boundaries or an inability to manage adolescence. Families often see the pattern between professional contacts. They notice the sequence, not just the episode.
The current London Assembly Police and Crime Committee investigation into grooming gangs in London is examining prevention, investigation, victim support and partnership working. Those questions cannot be answered by counting incidents alone. They require scrutiny of how signs were interpreted, whose account was believed, what agencies knew and whether anybody joined the fragments together.
Five questions to ask before calling it a choice1. Who benefits from what the child is doing?
Follow the benefit. Who receives the money, status, sexual access, information, transport, premises or new recruits? A child may be the most visible person in the chain while holding the least power.
2. What might happen if the child refuses?
Do not limit this to an explicit threat. Consider violence, humiliation, exposure of images, loss of protection, withdrawal of affection, debt, retaliation against family members or fear of police and services.
3. What threat, reward, debt or dependency is operating?
Map both pressure and attachment. Exploitation can be maintained by fear and by the child’s need for belonging. The reward does not cancel the coercion. It may be one of its mechanisms.
4. What information is held by other agencies or the family?
Ask what changes when the missing episode, school absence, injury, new phone, location, peer group and offence are placed on the same timeline. Identify who is responsible for building and updating that picture.
5. What would need to change for the child to feel safe enough to speak honestly?
Disclosure is affected by safety, trust and what the child believes will happen next. Consider who should speak with them, where, at what pace, and what practical protection can be offered before asking for more detail.
What a better professional response looks like
Greater curiosity must lead to different action, not simply more questions for the child.
A stronger response keeps a consistent trusted professional involved, avoids closing support because engagement is uneven, and builds a safety plan around the places and people connected to harm. It records facts clearly while separating observation from interpretation. It asks what can be disrupted around the adult, network or location, rather than placing the entire burden of safety on the child.
It also avoids repeated interrogation. The child should not have to produce a perfect disclosure for each new professional in order to remain eligible for belief. Information should move safely between agencies so that the child does not carry the system’s communication burden.
And it pays attention to language. Compare ‘she keeps choosing to return to him’ with ‘she has returned to contact with an adult about whom there are concerns regarding coercion and control’. The second formulation does not manufacture certainty. It keeps the safeguarding hypothesis visible and directs attention towards the adult’s behaviour.
Professional curiosity is not about removing all agency from children. It is about refusing to award adult freedom to a child whose choices may have been narrowed by fear, dependency and control.
The question we should leave withThe child who refuses help may be the child who has learned that help cannot protect them. The child who returns may be responding to a threat nobody has yet uncovered. The child who protects an adult may be protecting a sibling, a friend, their reputation or the only relationship they believe they have.
None of this means professionals should assume exploitation whenever a child takes a risk or commits an offence. It means we should be as curious about control as we are about conduct, and as interested in the adult creating the conditions as we are in the child reacting to them.
The question is not simply why the child keeps returning. The question is what the person exploiting them has made them believe will happen if they do not.
Evidence and further reading
Home Office: Criminal exploitation of children and vulnerable adults, county lines. Defines child criminal exploitation and describes coercion, blackmail, online entrapment and debt bondage.
Centre of expertise on child sexual abuse: Child sexual exploitation. Summarises research on control, coercion, apparent cooperation and the overlap between online and offline exploitation.
Centre of expertise on child sexual abuse: Identifying and responding to disclosures. Explains disclosure as a process that may be partial, delayed, behavioural or non-verbal.
The Children’s Society: Appropriate Language, Child Exploitation. Shows how victim-blaming language can affect professional thinking, disclosure and support.
Joint inspectorates: Multi-agency responses to serious youth violence. Reports on identification, information-sharing, trauma, exploitation and joined-up safeguarding practice.
Department for Education: Working Together to Safeguard Children 2026. Current statutory guidance for multi-agency work to help, support and protect children in England.
London Assembly: Grooming gangs in London investigation. Sets out the committee’s current focus on prevention, investigation, victim support and partnership working.
RSS Feed