Schools aren't rejecting AI. They're rejecting AI they can't account for.
A teacher has no way to tell a student who learned from a student who copied. So the whole category gets banned. That is a reasonable response to missing information — and the missing information is what OSQR produces.
Homework used to carry a second signal beyond the answer: it was weak evidence that a student had practiced. That signal is gone. What replaced it is suspicion — of every student, including the ones who did the work.
A ban does not restore the signal. It just moves the AI to a device nobody can see, where it does the whole assignment in ten seconds with no one watching.
Everything in this space watches or grades. Nothing vouches.
These are real tools doing real jobs. The point is what none of them can do.
Device monitoring
Flags what a student typed into an AI tool on a school-issued device.
Most homework happens at home, on a family phone or laptop the school does not manage.
AI detection
Scores a finished document for how machine-written it looks.
It judges the artifact after the fact, and a wrong call falls on the student.
Free study modes
Ask questions back instead of answering, in most major chatbots.
It is a setting the student controls. A student under deadline pressure turns it off.
All of it looks for evidence a student did something wrong. None of it produces evidence a student did something right.
A page the student prints and hands to their teacher.
After a homework session, a parent can turn the conversation into a one-page record of how the work actually got done. It is built from the full transcript, and the student can print it but cannot write or change a word of it — which is the only reason a teacher should believe it.
- How many minutes the student actually worked, with idle time excluded
- How many times they went back and forth with the tutor
- Where they got stuck, and what got them moving again
- Which misunderstandings came up, and whether they were resolved
- Whether the final answer was ever provided — stated plainly, either way
The line a teacher reads first
“Self-discovery was set to 8/10, a level the student cannot lower. The final answer was not provided — the student reached it themselves.”
And when an answer was given, the receipt says that instead. A receipt that made every student look diligent would be worthless to the teacher reading the second one.
A setting the student can't turn off.
Most chatbots now have a study mode that guides instead of answering. In every one of them, it is a toggle the student controls — which means it is off at 10pm the night before something is due.
In OSQR, a parent sets the level from 0 to 10 and it acts as a floor. A specific assignment or a teacher-built study guide can push it higher. Nothing the student does can push it lower.
“Can you read my kid's conversations?”
We can read them. We can't read them without leaving a permanent record you can see.
Every time someone at OSQR opens a child's conversations through our tools, it is written to a log the parent can open — who looked, when, and why — and it is written before they see a word. The entries are hash-chained, so one cannot be altered or removed without breaking every entry after it.
We are deliberately not claiming your child's data is mathematically unreadable to us. It isn't, and it can't be: a parent can read everything their child writes, and the safety checks that watch for a child in distress only work because something is reading. Encrypting that away would mean giving those up. So instead of a promise you would have to take on faith, you get a record you can check.
What we're asking a school for: nothing.
We are not asking a school to buy software, sign a data agreement, hand over a student roster, or endorse a vendor to families. Parents create and pay for their own child's account, exactly as they would for a tutor.
The only useful thing a school can offer is clarity: tell families what is and isn't acceptable, so a student using AI honestly isn't treated the same as a student using it to skip the work. That costs nothing, and it is worth more to families than any product is.
If it helps, we wrote a policy template you can adopt, edit, or ignore — free, no signup. It works whether or not you have ever heard of us again.
The honest limits
Can a student get around it?
A determined student can open a different chatbot in another tab. Nothing stops that, and any vendor who tells you otherwise is selling you something. What a receipt does is make the honest path provable — so a student who did the work can show it, instead of being suspected alongside everyone else.
Is the receipt generated by AI?
Yes. A model reads the finished transcript and reports what happened in it. It is instructed never to flatter — if a student coasted or kept demanding answers, the summary says so. The full conversation stays available to the parent, so any receipt can be checked against the thing it describes.
Does the school have to buy anything?
No. Parents create and pay for their own child’s account. There is no school contract, no student roster, and no student data flowing from the school to us.
Who can see a child’s conversations?
The parent or guardian who created the account, in full. Child accounts are not encrypted away from their guardian — that is deliberate, because guardian oversight is the entire safety model. Data is encrypted at rest and isolated per account, and any time someone at OSQR opens a child’s conversations, the parent sees it in their account.
We're not selling AI adoption. We're selling accountability.
If a school never allows a single AI tool, that is a legitimate choice. Students will still use them. The question worth answering is whether the ones who are honest about it have any way to prove it.