Moderation Policy
Overview
Our moderation policy is designed to help our users avoid potential legal risk, keep us compliant with the Online Safety Act (OSA) & UK GDPR, and to create a useful and enjoyable community for everyone.
But... Moderation is... Complicated…
For a good example of the complexity, see below the challenge of moderating (and, more importantly, the reason for its moderation) the use of specific job titles (i.e. "The D.T. Teacher") in reviews.
We need to balance freedom of expression with legal and compliance risk, and we want to treat all of our users as grown ups. To this end:
we use ‘Just in Time' moderation blocks and hints throughout your journey to help you understand where you might be going wrong, but...
we also make it clear where your responsibilities and liabilities sit in our Terms.
We employ three levels of moderation:
Just-in-time. These are real-time "audits" of the content you attempt to publish that flags OSA, UK GDPR, or significant Terms violations. These are automated processes that either block your submission until the violation is resolved, or collect the required consent from you to store and process that data.
Manual review. Certain content may be held pending manual review before it is published, or may be manually reviewed post-publication.
Our complaints process. If anybody catches something we've missed, we have a robust Complaints Procedure and takedown process.
But the most effective level of moderation will always be ‘Level 0’.
Level 0 is you, the user, understanding our Terms and Community Standards fully and moderating yourself before you attempt to submit any information.
Thank you for your support!
Why we moderate
We moderate for three reasons:
To help protect you, our users, from opening yourselves up to legal risk.
To protect us, Indeparent, from legal liability.
To create a healthy community, in line with our Values.
Protecting you
Under Section 5 of the Defamation Act 2013 and the Electronic Commerce (EC Directive) Regulations 2002, Indeparent acts as an "intermediary" or "operator of a website hosting user-generated content".
This means that when you submit content, having accepted our Terms of Service (which incorporates other documents including our Acceptable Use Policy), you acknowledge that you, the user, are legally liable for the content you post and that you are ultimately liable for any legal disputes that may arise.
The aim of our moderation policy is to detect and intercept potentially problematic content before it is even stored on our server with our ‘Just in Time’ audits, and to catch whatever else we can in pre- and post-publication moderation.
However, there are certain risks that we cannot practically detect. A great example of this is defamation (see what is defamation for more detail).
A comment like “The Head lied to us” certainly has a high risk of being defamatory.
However, we, Indeparent, could never know for sure. You may have incontrovertible evidence that The Head did indeed lie to you (it may even be in the public record if you have proven in a court that you were lied to), in which case there is a strong legal argument that the comment is not defamatory and that you are exercising your freedom of speech in sharing it.
Given that it is not possible to assess the truth of claims, we do not attempt to moderate your content in these respects.
Therefore, whilst we attempt to intercept the most clear-cut and unambiguous risks, you must remain alert to the risks of what you post.
Protecting ourselves
We, Indeparent, also risk legal bother if we don’t take our data protection and other regulatory requirements seriously (which is why we do!).
We are required to:
Comply with the Online Safety Act (OSA): We have a statutory duty to proactively assess risks, block priority illegal content (such as child exploitation or extreme violence), and ensure children are protected from encountering harmful material.
Uphold the UK GDPR: We must protect personal privacy by ensuring we only process highly sensitive "Special Category Information" (such as health data) with valid legal consent, and by preventing the unlawful identification of individuals.
Adhere to the Defamation Act 2013: To maintain our legal protection as a platform "intermediary" (so we are not held liable for the raw reviews you write), we are required to follow a strict 48-hour Notice and Takedown procedure whenever we receive a formal complaint about a potentially defamatory statement.
This document outlines the steps we take to meet these requirements.
Creating a healthy community
There are many things we audit that are not legal requirements; we’ve just made the choice to prohibit certain things from our platform.
Severe profanity, obscenity, or vulgarity is a great example. We don’t prohibit swearing because we have to, but because we want to.
It may seem prudish, but it’s a choice we’ve made. It’s not a moral objection (we swear like dockers in our private lives), but we want to create a platform that parents would feel comfortable browsing with their kids.
We’ve made our reviews anonymous for good reasons, but the veil of anonymity can lead people to use the kind of language that they wouldn’t feel comfortable using if their name were attached to it, so we’ve made the call to pro-actively block such language as best we can.
What we moderate
Online Safety Act (OSA) violations
The OSA is a serious bit of legislation. Like most legislation that has the best of intentions (which it does), it can introduce restrictions that don’t seem commonsensical at first glance. This is why we’re taking the time to explain the impact of the OSA on our moderation policy.
With regard to the OSA, we moderate for:
Child Sexual Abuse or Exploitation (CSAE). Content promoting, depicting, describing, or grooming child sexual abuse/exploitation.
Encouragement or Instruction of Suicide or Self-Harm. Content that encourages, glorifies, instructs, or provides methods for suicide, suicide attempts, or self-harm.
Terrorism, Extremism or Incitement to Violence. Content inciting to terrorism, glorifying terrorist groups/acts, or instructing on weapons/explosives. Direct, explicit threats of severe physical violence or death.
Incitement to Hatred & Hate Speech. Extreme hate speech that systematically promotes hatred, severe hostility, or violence against protected groups based on race, ethnicity, religion, disability, gender, sex, or sexual orientation.
Severe Harassment, Cyberbullying & Malicious Threats. Extreme targeted abuse, malicious communications designed to cause severe emotional distress/panic, or doxxing (publishing private personal data like home addresses, personal phone numbers, or private emails of staff/students with malicious intent).
Other Serious Illegal Activity. Encouraging, facilitating, or listing instructions for serious illegal crimes (e.g., selling illegal Class A drugs, weapons trading, financial scams, hacking, or credit card fraud).
The above categories are pretty-much verbatim from the official legislation.
We have smart systems that attempt to parse the content for violations and block them before they even hit our database, but no system will ever be 100%.
False positives. We flag something as a violation that is perfectly reasonable. You will usually be able to clear the block by slightly changing the wording of the offending snippet (which we will flag to you). Please let us know if you encounter a false positive so that we can review our processes.
False negatives. Some content passes our audit that is actually in violation of the OSA. Here we will remind you that you remain liable for the content you post, subject to our Terms. Our failure to intercept the content before it was stored or published will not be an effective defence for you.
One really complex thing to moderate against The OSA is the use of specific job titles.
“The D.T. teacher is incompetent” could very easily fall foul. In a small school there may be only one D.T. teacher, and so this statement is a public claim of incompetence against an easily identifiable individual.
This is why we prohibit the naming of any specific individual, or even the use of specific job titles (with a carve out for “executive roles”, see below for more).
Special Category Information
Special Category Information is defined under Article 9 of the UK GDPR as any information that falls into the following categories:
Racial or ethnic origin (e.g., specific race, national/ethnic background of any individual).
Political opinions (e.g., personal voting preferences, specific political views of any individual).
Religious or philosophical beliefs (e.g., personal faith of any individual).
Trade union membership (e.g., member of NEU, striking).
Genetic data (e.g., DNA analysis, genetic testing).
Biometric data (e.g., fingerprints, iris scans, facial recognition).
Data concerning health (e.g., specific medical conditions, disability, neurodivergence, autism, adhd, dyslexia, EHCP, SEN/SEND, etc.).
Data concerning a person's sex life.
Data concerning a person's sexual orientation.
In the case of Special Category Information, it’s not that it’s prohibited, it’s that we require special consent from you to store and process this information.
Our audit parses your content in temporary memory and then discards it, ensuring that it is never stored permanently on our server and is only processed for the purposes of compliance auditing. This temporary processing is compliant because it relies on two lawful bases: Article 6(1)(c) (Legal Obligation) to satisfy our duties under the Online Safety Act, and Article 6(1)(f) (Legitimate Interests) to enforce our Acceptable Use Policy.
If our audit detects Special Category Information, then this will be flagged to you and you will be asked to provide explicit, informed consent for us to store, process and publish this information (subject to our Terms and Privacy Policy).
Note that you have the right, at any time, to withdraw this consent by deleting your review, blog comment, or your entire profile or account.
However, it’s important to note that you will only be asked to provide this consent if it pertains to yourself or your own child (or a child for whom you have legal guardianship). If the information appears to relate to an individual for whom you cannot legally provide the consent for it to be stored, your content will be blocked until it is removed.
This is why it’s very important that you faithfully characterise “your relationship with the school” at the beginning of your review. If you select “Parent/Legal Guardian" or “Ex-Parent/Legal Guardian" then you will be able to consent to your information and for that of your child to be stored. Otherwise, you will only be able to consent to your own information being stored and processed.
This may seem strict, but we need to be conscious of likely scenarios. It’s perfectly possible that the Nanny of a child will create a review and add a comment such as “my child has ADHD, and they cater really well for her”. The Nanny DOES NOT have the legal basis to consent to that Special Category information (SEND status) to be stored and processed by us.
Terms violations
We also parse your content “just in time” to audit it for obvious violations of our Terms (including our Acceptable Use Policy).
As per all of these moderation steps, a “pass” of this audit does not imply that your content is free from violations. No process is 100%, and there will be false-positives and false-negatives.
Specifically, we attempt to scan for:
Swearing. This one’s really tricky, because there’s a really hard to draw line between “proper swearing” and “flowery language”. There’s an awful lot of slang, misspellings and esoteric cuss words out there, and we’ll never catch them all. We maintain a very long (and quite amusing) list of “the big ones”, and check your content against that.
Doxxing. This is actually redundant at this stage because of our OSA audit but, just for good measure, we have also added screening here.
Promotion. Promotion of 3rd party businesses is prohibited by our Acceptable Use Policy.
Naming of individuals. Again, this is technically redundant in this step because it will be caught in the OSA/Special Category audits.
Use of specific job titles. For the reasons given above, this is prohibited within our Acceptable Use Policy. Under the "Journalism Exemption" (Schedule 2, Part 5 of the Data Protection Act 2018), “executive” titles, such as “The Head”, “The Bursar”, “Chairman of The Governors” are permitted on the basis that their roles are in the public record.
How we moderate
Time for some technical detail, which will help you understand the process and how your data is handled.
“Just in Time” processing
When you enter a piece of information into a text box on our site, it is blind to us until you save it. Technically speaking, the data sits on “the client side”.
When you hit save, you will see that the button says “Auditing”. At this stage, your data will be passed to our server for processing, but it will only ever be held in temporary memory specifically for the purpose of the compliance audit and it will be discarded as soon as the audit is complete. This is compliant data processing as noted above in this document.
If your content passes the audit, then great news - your information will be saved to our database and you will proceed directly to the next step (reminder: this does not imply that your content is free from any legal risk).
If your content fails the audit, then you will see a notification telling you why, what specific piece of content was flagged, and what to do next.
If it is a Special Category Information alert, then you will be asked to provide consent for us to store and process this data (subject to our Terms and Privacy Policy) if you have the legal right to provide such consent.
Otherwise, your progress will be blocked until you remove the non-compliant information.
Our Just in Time audits have three stages:
Online Safety Act (OSA). Any failures are a hard block, and no further audits are run until the content is updated or removed.
Special Category Information (SCI). If the OSA audit passes, then the SCI audit takes place. Any failures will either result in a request for consent to store the information, or a hard block, depending on the context.
Terms. If the SCI audit passes, then the final step is the Terms audit. Given that we only attempt to audit for the most obvious and unambiguous violations, this is always a hard block. Progress is only possible once the offending content has been removed or edited.
Manual review
Certain content, subject to our risk assessment, will be held for manual review before it can be published. We will audit using the same criteria as the Just in Time automated process, but with the subjective eye that only a real-life human can bring to the party (for now…).
Other content will be published immediately, but may be subject to post-publication manual review.
Complaints and takedown procedure
We think the above is something like best practice for moderation. We’ve tested it thoroughly, and we’re pretty proud of the results. However, no process will ever be 100%. That’s not an admission of failure, just reality.
Our aim is to flag and intercept anything critical before it is published (and, ideally, before it’s ever saved to our database), but our complaints and takedown procedure is there as a backstop to deal with anything that slips through.
If the complaint points to a violation of our Terms (illegal activity, threats of violence, etc.), then we reserve the right to immediately remove that content.
For other, more ambiguous, legal complaints that we are not in a position to adjudicate (accusation of defamation, etc), we have our formal complaints procedure.
You can read the policy for more details, but the TL;DR is:
Anyone can raise a formal complaint about any piece of information entered by a user.
We forward that complaint to the user (who is liable for the content they post).
If they do not respond within 48 hours, we reserve the right to remove the content.
If they respond and request for the content to be taken down, we will help them do that.
If they stand by their content and refuse its removal, then the content remains up and the legal dispute is between the user that posted the content and the person with the complaint.
Key takeaways
The policy aims to balance freedom of expression with legal compliance (OSA and UK GDPR) and creating a healthy community.
Moderation operates on four levels: Level 0 (User Self-Moderation), Just-in-Time automated audits, Manual Review, and a Complaints Process.
Users are legally liable for the content they post, and the platform's failure to intercept non-compliant content is not a defense.
Hard blocks are applied for Online Safety Act (OSA) violations (e.g., CSAE, hate speech, incitement to violence/self-harm) and significant Terms violations (e.g., severe swearing, naming individuals, specific non-executive job titles).
Special Category Information (e.g., health, race, political opinions) requires explicit, informed consent from the legally responsible party (self or own child) for storage and processing; otherwise, the content is blocked.
The platform does not moderate for the truth of claims like defamation; users must remain alert to this risk.
The Complaints and Takedown Procedure is the backstop: for ambiguous legal complaints, the user is notified and given 48 hours to respond before the content may be removed.