How does your AI learn to differentiate between pictures and audio and video?
In AdHive we have specialists, who teach the AI according to the cases of advertisers and they use the TrainTool.
What is TrainTool? I mean where can I learn more about this technology?
It's a program or a tool that is used to train AIs, you can read more about it here[1]. i haven't yet tried/used the traintool, but soon i'm going to try it.

[1]
https://medium.com/adhive/ai-department-update-traintool-release-2-0-4c26780b2debYeah, you are absolutely correct

Glad to see that you take huge interest in the development of the project! Thanks for the support!