Safety has been our top priority as we designed Eldera. We promise not to sell your data, we background check each mentor, and we are committed to keep parents in the loop. Every Eldera mentor passes a live human screen, all sessions are recorded, and AI-assisted reviews keep everyone safe. Here is exactly how it works.
Most platforms initiate a safety measure only after something goes wrong. We built Eldera with safety in mind to prevent problems before they start.
Too many tech companies are reactive, with new guardrails that appear only after a crisis. We built Eldera the other way around, with built-in proactive safety processes. From our first sessions, every call is captured on video and kept in our archive. Mentors and young people are matched from different parts of the country to keep conversations on the platform. Every mentor passes a background check. The AI Chaperone came next. It is our proprietary tool that reviews sessions and flags for human review potential violations of our Code of Conduct. We made these choices not in response to a problem, but to prevent one.
As I direct the work on these and other safety measures, I am guided by the security I want for my own family.
CL
Real numbers from a platform six years in operation. Stated plainly, never inflated.
No single safeguard succeeds on its own. These five work together to provide layers of protection for children, mentors, and the full Eldera community, from the moment a mentor applies to long after a relationship has begun.
No exceptions, no shortcuts. Each mentor passes a criminal background check, run by our screening partner, before we match that mentor with a child. Eldera pays for the check ourselves. It is part of joining, and never a step anyone can skip.
✓ Paid for by ElderaA background check is a record. A conversation is a judgment. Before a mentor is approved, we require more than just paperwork, we arrange a live orientation with our team. We get to know who you are, walk through our expectations and safety practices, and make a human decision about who belongs in our community. Many applicants do not make it past this stage, and that is the point.
✓ A human screen, not a checkboxYou are introduced to your child's mentor before they ever meet, and you see the mentor's full Eldera profile before anything is confirmed. For children under 13, the account is yours to manage and you are present for the first session. After sessions, you receive a summary of what was discussed, so you stay close to the relationship without standing in the middle of it.
✓ You see the relationship, alwaysSessions take place on the Eldera platform, and each one is recorded on video and kept in our secure archive. Nothing happens off the record or out of view. The recording enables extra layers of safety: it is what the AI Chaperone reviews, what a human can check if a question is raised, and the reason a conversation can never quietly disappear. The recordings protect mentors, too, giving a complete account of what was said if any conduct is ever questioned.
✓ Nothing happens out of viewOur proprietary AI Chaperone reviews the video and transcript of every session for signs of risk, across more than a dozen safety categories. It runs quietly in the background and never joins or interrupts the conversation. When it flags something, a human member of our team reviews it. The AI never acts alone. More on how it works below.
✓ Human-reviewed, never automated aloneThe AI Chaperone is a safeguard, not a participant. It does not talk to your child, steer the conversation, or come between the child and mentor. It reviews the recording and transcript of each session afterward, looking only for signs that someone might be at risk.
When it notices something that needs a closer look, it does not decide anything on its own. It raises a flag for a human member of our team, who reviews the flag and decides what should happen. The technology finds; a person judges.
That division is deliberate. AI is fast and tireless and good at noticing patterns across thousands of conversations. It is not who we trust to make a decision about a child. That is always a human being.
The video and text transcript of every session, after the conversation has happened. It works from the record, never as a live listener in the room.
Signs of risk across more than a dozen safety categories, with a catch-all for anything that does not fit a named one:
A human member of the Eldera team reviews it. The AI never takes action on its own, and no concern is ever closed by automation alone.
Good guardrails are not about locking a relationship down. They are about keeping it open, visible, and always connected to the parent. Here is where the lines sit.
Sessions happen on our platform, where they are recorded and reviewed. That is what keeps the relationship observable and safe, rather than moving into private channels no one can see.
The parent is always the center of the relationship. Mentors agree in the Eldera code of conduct to never arrange anything with a child directly, online or in person. That means if families and mentors connect beyond a session, it happens with the parent leading and fully aware.
We deliberately match mentors and mentees from different parts of the country. It keeps the relationship rooted in the platform and on screen, where it can be seen and supported.
Mentors are guided by a code of conduct. Asking a child to keep something from their parent is a violation of that code. Warmth is welcome and secrecy is not.
A card or a care package between people who have grown close is part of what makes this community special. What is never allowed is a mentor asking a child for money, or any private financial arrangement that goes around the parent.
For children under 13, we obtain verifiable parental consent before collecting any information, in line with COPPA. The account is parent-managed throughout.
Prevention is most of safety, but it is not all of it. The most important thing a safety system can tell you is what happens when something does not feel right. Here is ours.
If anything about a session, a message, or a mentor ever feels off, you reach out to our team. You do not need to be certain, and you do not need proof. A feeling that something is not right is reason enough.
A real member of the Eldera team looks into it. Because every session is recorded, we can review exactly what happened rather than guess. The memory of a participant is not the only account of events; the record is.
Depending on what we find, we may pause the match, remove a mentor, or escalate to the proper authorities. Where there is any sign of harm to a child, we report it to law enforcement and the appropriate child-safety bodies. We do not handle these quietly or on our own.
Not every concern is a safety issue. Sometimes a mentor and a child just do not click. That is normal, and it is never awkward. Tell us, and we will find a new match. No child is ever stuck in a relationship that is not working for them.
To raise a concern about a mentor, a session, or anything else, contact our team at dana@eldera.ai. We read every message.
We do not claim a perfect record, because no honest safety system can. Our incident rate is under three in ten thousand sessions. It is not zero, and we will never pretend it is. What we can promise is that nothing happens out of view, every concern reaches a person, and we built it this way on purpose, from the beginning.
If you have read this far, you care about the same things we do. We would be glad to introduce your child to a mentor who has earned their place in this community.
Find My Child a Mentor →