We spend enormous amounts of time talking about how to make the internet safer for users. We talk about child protection, content moderation, online harms, platform accountability and Trust and Safety. We talk about how quickly harmful content should be removed, how platforms should respond to reports, how artificial intelligence can detect abuse, and how policies can better protect children and adults online.
But for a long time the safety of the people doing the protecting has hardly been part of the conversation.
Behind every removed piece of harmful content, every escalated report, every child safety investigation and every decision about whether something violates a platform's policy, there is a person looking at the screen, making a judgement, deciding whether something is dangerous, illegal, exploitative, abusive or simply harmful.
We call them content moderators, child safety investigators, policy enforcement specialists, escalation team members, digital forensics professionals, abuse response specialists and Trust & Safety analysts. Their job titles may differ, but many of them are exposed, repeatedly or routinely, to the consequences of online harm in order to protect the rest of us from it.
The people behind the safety layer
The public often experiences Trust & Safety as a set of outcomes, a post disappears, an account is suspended, a report is investigated, a child is identified as being at risk, a scam is detected, an abusive network is disrupted, a piece of illegal content is removed. We see the outcome. We rarely see the human process behind it.
We don't see the practitioner who may have reviewed hundreds of pieces of content before lunch.
We don't see the investigator who has to repeatedly examine evidence of the exploitation of children.
We don't see the analyst who spends hours distinguishing between harmful content that needs immediate escalation and content that falls outside a particular policy threshold.
We don't see the person who closes their laptop at the end of a shift and is expected to transition immediately back into ordinary life.
That invisibility is part of the problem.
Trust & Safety work is often discussed as a technical, legal or operational function. And it is all of those things. But it is also human work performed under conditions that can carry significant psychological demands.
Research involving content moderators has found a relationship between the frequency of exposure to distressing content and psychological distress and secondary trauma. The same study found that supportive colleagues and a sense that the work was important could help mitigate that relationship.
Research involving digital forensics professionals who work with child sexual abuse material tells a similar story. Participants described difficulties decompressing from the work and reported experiences associated with compassion fatigue, secondary traumatic stress and burnout.
These findings should not be interpreted to mean that every person working in Trust & Safety will experience psychological harm but should challenge the assumption that people can simply be exposed to difficult material indefinitely because it is their job.
Resilience is not immunity
There is a tendency to use the word resilience as though it means the ability to withstand anything. A resilient practitioner is sometimes imagined as someone who can see difficult material, make difficult decisions, absorb difficult conversations and simply carry on.
But that should not be the organisational definition of resilience. If resilience becomes synonymous with “how much can this person tolerate before they break?”, we have designed the wrong measurement.
The World Health Organization identifies excessive workloads, understaffing, long or inflexible working hours, low job control, limited workplace support and other psychosocial risks as factors that can threaten mental health at work. However, it is evident that risk does not necessarily come from exposure alone.
Consider the difference between two practitioners.
One reviews difficult material for a limited period, has meaningful rotation, takes protected breaks, has supportive colleagues, knows how to access confidential professional support and has some control over their workflow.
Another reviews similarly difficult material, but works long shifts, faces constant queues, is measured primarily by speed, skips breaks because of a Service Level Agreement (SLA), rarely rotates away from high-severity work and worries that admitting distress will affect how they are perceived.
The content they are exposed to is similar but the working conditions are not. We therefore need to start questioning how resilient the systems the practitioners are working in are.
When the work follows you home
One of the particularly difficult characteristics of some T&S roles is that the practitioner may be required to make repeated contact with other people's trauma. A digital forensics analyst may encounter evidence of child sexual abuse, a child safety investigator may spend hours working through exploitation cases, a content moderator may encounter violence, abuse, self-harm, exploitation or other disturbing material and an escalation specialist may repeatedly deal with the most severe cases because they have the experience required to handle them.
Research on digital forensics professionals has participants describing intrusive thoughts, disturbed sleep, hypervigilance, mistrust and difficulty disconnecting as a result of what they had encountered.
Sometimes it is easier to conclude that they are simply stressed. But if we reduce every problem to “stress”, we can also fail to recognise burnout, secondary traumatic stress, vicarious trauma and compassion fatigue which are related concepts, but they are not interchangeable. Understanding the difference matters because the appropriate response may also be different.
The dangerous success story
We also need to talk about the hidden trap of the high performer. These are the employees who always take the hardest cases, volunteer for overtime, and never complain. They are celebrated as "strong" and always claim to be fine.
But their success can mean they are carrying a disproportionate share of the team's psychological burden. The critical question then becomes “Are organizations rewarding true resilience, or are they just rewarding the ability to tolerate an unhealthy culture?"
When we continuously assign the heaviest workloads to the most reliable people, their excellence becomes a hazard. If they eventually burn out, we cannot blame it on a failure of personal self-care but on the system that needed them to carry so much of the burden in the first place.
Culture matters too
Telling someone to “speak up if you're struggling” assumes that speaking up carries little social or professional risk. That assumption cannot be universalised.
Research on mental health help-seeking has identified stigma, self-reliance, difficulty disclosing distress, low mental-health literacy and concerns around available services among the barriers people can face when seeking psychological support. That means a wellness programme cannot simply be designed around the assumption that everyone understands mental health in the same way or feels equally comfortable discussing it.
A culturally competent wellbeing programme has to consider not only what support is available, but whether people actually feel safe enough to use it.
From individual resilience to organisational responsibility
There is an understandable desire to give practitioners tools; mindfulness, breathing exercises, counselling, peer support, mental health literacy, physical activity and time away from the screen. But they cannot be the entire strategy.
The WHO's approach to mental health at work explicitly includes organisational interventions alongside manager training, worker training and individual interventions.
If a team is understaffed, giving people a mindfulness session does not resolve understaffing.
If practitioners are routinely working excessive hours, teaching them breathing exercises does not change the hours.
If the most traumatic cases are repeatedly assigned to the same people, a wellness webinar does not redistribute exposure.
If employees are afraid to disclose distress because they don't trust the confidentiality of the support system, simply announcing that counselling is available does not create psychological safety.
Wellness cannot become a substitute for fixing the workplace. The responsibility for practitioner wellbeing therefore sits at several levels.
There is the individual, who needs knowledge, coping strategies and access to appropriate support.
There is the team, which needs supportive managers, psychologically safe conversations and healthy peer relationships.
And there is the organisation, which has responsibility for workload, staffing, shift design, exposure management, workflow, technology, targets and the culture surrounding help-seeking.
The strongest systems work at all three levels.
Protecting the protectors
We have become increasingly sophisticated about protecting users from online harm.
Perhaps it is time to become equally sophisticated about protecting the people who carry out that work. We should not aim to create practitioners who can tolerate more harm, but to create systems that require practitioners to absorb less unnecessary harm while giving them meaningful support when exposure is unavoidable.
This is where this conversation begins...
In the next blog, we look at what repeated exposure to online harm can do to the people behind the safety layer, and what organisations need to understand before they can meaningfully protect them.
Because protecting the internet should never require us to sacrifice the people protecting it.
About the Author
Athena Mwarwakamori Morgan is the Founder of Mindful Clicks Africa, where she advises schools, organisations, and policymakers on digital governance, child online safety, data protection, and responsible technology use. She has worked across Africa on child protection, technology policy, and safeguarding, helping institutions create safer and more resilient digital environments.