Question about ML4kids GDPR compliance

[jurij marhold]

Hello,

I would like to know whether the ML4kids service is compliant with the European data protection regulation (GDPR). Since we plan to use it in the EU, we would like to understand:

• whether all data processing procedures are adapted to GDPR requirements,
• if not, what steps we as users need to follow to avoid any GDPR violations,
• and whether there are specific guidelines for schools, parents, or children in the EU when using ML4kids.

I would appreciate any clarification or a link to official documentation confirming the service’s compliance.

[Dale Lane]

Details on this can be found at https://machinelearningforkids.co.uk/help 

Specifically, the sections on:

* What information is stored about users?

* What happens to training data created by students?

In summary, every part and feature of the tool can be used anonymously. The "Try it now" option on the login screen lets you use the site for four hours (intended to be long enough for any typical school lesson or code club session) without creating any user accounts or providing your name or contact details.

If you choose to create an account because you want to store projects for longer than four hours, then I ask for an email address for the teacher. This is solely to allow you to reset your password if you lose/forget it. It isn't used for any other reason. 

Teachers are responsible for creating and managing user accounts for students. As long as they use anonymous usernames for students as I recommend, this means that I don't store any names or contact details for student accounts. 

When you create projects, you can choose where to store the data for your project - in the cloud (that is to say, on my server - further technical specifics about what data is stored and exactly where it is stored can be found at the link above) or on your own computer. 

If you store projects only on your own computer, then you naturally remain in control of your own data. (The only exception to this is for models to recognise text, which are trained by submitting your training examples to an IBM Watson cloud service - more details on this can be found at the link above. Other model types - images, numbers, sound, etc. - are trained on your own computer, so if you store projects only your own computer the training examples are stored only on your own computer).   

Finally, I should add one piece of context - which is that Machine Learning for Kids is a website that I run as a personal/side project. While I'm very happy to provide as much technical information as you need to satisfy your organisational requirements, please bear in mind that I have no legal or compliance team helping me with this. The benefit of this is that I can make it freely available with no fees or adverts. The downside is that the capabilities of the site are limited to what I can do or what is kindly offered by more knowledgeable volunteers such as yourself. 

Kind regards

D