|
|
|
|
|
|
by mbrundle
439 days ago
|
|
|
I had an idea to screen the YouTube videos that kids watch through a suitable LLM to assess whether the content was child friendly. There’s an emerging issue with insidious messaging / themes creeping into child-focused content that looks completely innocent on the surface, and I wondered if the current generation of LLMs can pick up on this. (Initial tests seem encouraging.) |
|
|