TalkCampus is not
an AI product.
We believe in the power of real people connecting with real people — now more than ever.
TalkCampus is a peer support community, powered by human moderation. AI is used internally as a tool that helps our Trust and Safety team work more efficiently — nothing more. There are no chatbots. No automated responses. No AI speaking to your students.
Fact
AI does not talk to users
Every message is written by a real student. AI has zero presence in conversations.
Fact
AI does not moderate our platform
AI flags. Humans decide. Every moderation action requires a person.
Fact
AI is an internal tool only
Students never see or interact with AI. It runs behind the scenes for staff.
What AI does — and does not — do
There is no ambiguity in how we use AI today, and none in our future roadmap. While we follow developments in AI-assisted mental health with genuine interest, AI-led interventions in peer support remain largely unevidenced at scale. They also introduce a set of clinical safety, liability, and student trust considerations that we are not willing to take lightly. Our position is clear: humans support humans on TalkCampus.
AI does not talk to users
Students talk to each other. AI has no presence in conversations, no chatbot in the peer support feed, and no generated responses. Every message a student reads was written by another student.
AI does not moderate our platform
AI supports human-led moderation across our clinical and Trust and Safety teams. Every moderation decision is made by a trained professional.
AI supports human teams only
AI is an internal tool that helps Trust & Safety staff prioritise their queue. It runs in the background, invisible to students, surfacing signals so humans can act faster.
Clinicians moderate clinical risk. Not everything.
Clinicians are trained to assess clinical risk — not to identify coded language, spot grooming patterns, or recognise evolving slang used to evade filters. Mixing those roles doesn't make either better. Our model deliberately separates them because specialisation makes both safer.
What AI actually does on TalkCampus
Four limited, well-defined functions — each one ending with a human in the loop.
Content classification at creation
When a student publishes a post, multiple AI models score it for potential harm. Posts that exceed a threshold are placed into the human moderation queue. The post is not removed or hidden — a trained moderator reviews it and decides what, if anything, happens next.
Translation support
We use AI-powered translation to help students communicate across languages. The model is selected for contextual accuracy — not just literal word matching — so meaning, tone, and nuance carry across. This lets students support each other in their own language while keeping the community connected.
Classification to support moderation
AI assigns risk scores that help our human Trust and Safety team members identify signals across legacy posts and surface content that could be helpful when making decisions. This provides additional context that traditional clinical and Trust and Safety teams struggle to access when looking at individual posts in isolation.
Image safety
TalkCampus uses AWS Rekognition to automatically scan images before they are shared on the platform. This allows us to detect and block harmful visual content — including child sexual abuse material (CSAM), graphic violence, and explicit imagery — before it ever reaches another student. No harmful image makes it through unreviewed, and any detection triggers an immediate escalation to our Trust and Safety team.
What AI does not do
These are not caveats or edge cases. AI on TalkCampus has defined limits that cannot be overridden.
Diagnose, treat, or provide clinical advice
Replace or simulate a healthcare professional
Interact with users in any way — no chatbot, no generated messages, no AI responses
Make autonomous moderation decisions — every action requires a human
Provide therapeutic interventions of any kind
Profile, train, or learn from student data
Replace the specialist judgment of either clinicians or Trust and Safety professionals — both roles exist because neither can fully substitute for the other
Questions about AI on TalkCampus
Common questions from procurement teams, regulators, and student services.
No. TalkCampus is a peer support platform where students talk to each other. AI is used internally as an operational tool to help human moderators prioritise their work. Students never interact with AI.
No. Every message, reply, and conversation on TalkCampus is written by a real person. AI has no presence in the peer support feed and does not generate any content that students see.
No. AI flags content for human review and assigns a risk score to help moderators prioritise their queue. It does not remove posts, issue warnings, ban users, or take any enforcement action. Every moderation decision is made by a trained Trust & Safety professional.
No. TalkCampus is a peer support community, not a clinical or therapeutic tool. AI is used solely as an internal content moderation aid — it does not diagnose, treat, advise, or interact with users. This use case does not fall within the scope of medical device regulations under MHRA, FDA 21 CFR, or EU MDR frameworks.
We use multiple independent models from different providers running in parallel. This multi-model approach ensures no single vendor has sole authority and reduces the risk of false positives or false negatives. The models classify content — they do not generate responses or interact with users.
AI does not interact with students, so there is nothing to opt out of. AI operates entirely behind the scenes as an internal tool for the moderation team. Students experience a human-moderated peer support community.
Because clinicians are not trained for every type of harm that appears on a peer support platform. Identifying coded language, grooming patterns, predatory behaviour, and harmful slang requires a different skill set to clinical risk assessment. Our clinical team handle clinical risk. Our Trust and Safety team handle everything else. Separating those responsibilities makes both functions more effective — not less.
See how TalkCampus works for your institution
Book a demo to walk through our peer support platform, human-led moderation, safeguarding pathways, and reporting. We will answer every question about how AI is — and is not — used.