Service moderation for communication?

Question

Chats

Hello.
We have: an amateur project, developed in the evenings in your free time, is a service for meeting and communicating with a random interlocutor, which also allows you to share photos and other content.
In my head at the idea stage: a cheerful cozy chat talking about pink ponies and sharing photos of my cats.
In fact: a hotbed of pornography, tons of male penises, half of the dialogues about you know what.
At a certain stage in the development of the project, users began to receive complaints about their former interlocutors, complaints are not isolated. Appointed a moderator who fled after 2 days, according to him, even in the men's bath, he did not see so much ...
The question is: who is to blame, in terms of Russian law, if my users produce and exchange what is actually child porn (yes, they complain about girls too)? How to stop all this?
How, again, from the point of view of legislation, should a potential moderator of such a service act?
Block users, delete content, sprinkle holy water on the monitor and in your eyes, write statements to the relevant authorities?
Is the moderator of such a service a delinquent? He sees the content, it is even downloaded to his hard drive in the form of a browser cache...

Answer 1

When working with user-generated content, the main thing is to quickly process incoming complaints and communicate politely with their authors in order to make it clear (verbally) that you are on the side of the law and demonstrate a willingness to fight content that causes their outrage.
According to the Criminal Code, the distributors (content authors) are obviously guilty. But if the administration of the resource does not (adequately) respond to complaints about the content, then further complaints will fly to higher authorities, which will lead to the blocking of the resource.
Remove content and respond to all complaints and official requests. This is the main thing, and in 99% of cases this is enough.
The author's ban is a measure to facilitate the work of the moderator (less dirt will have to be cleaned up).
No. If he himself does not post such content.
But if he does not respond to complaints, then he substitutes the resource for blocking.
Naturally. That's his job.
If his house is searched (for some other reason), and he cannot intelligibly explain that he had no intention (a necessary condition of guilt) to save this content to disk, then they can try to sew an article of the Criminal Code on him (everything depends from lawyers and from his willingness to cooperate). But this is a purely hypothetical scenario; in reality, such cases are not yet known.

Answer 2

Permoban
Hellban
Image recognition
Prohibition of image posting
Binding to the phone, vk, fb, ok, ...
Paid image posting
Who is to blame - the one who is extreme and who can be reached.
If your person is within the reach of organs, then you are

Answer 3

To automate the detection of textual content, I can recommend using
https://pymorphy2.readthedocs.io/en/latest/user/gu... to bring all the words in the message into normal form and then feed the texts into https://pypi.python. org/pypi/redisbayes/0.1.3 to train the classifier.
This is how I did the spam protection on the classifieds site. After some training of the classifier, he himself could distinguish an ad from one category from an ad from another category in the text based on the words used. And he could also determine the reason why ads should be banned. (there were several reasons "financial pyramids" "drug advertising" and something else.)