also, 4o's image generation seems to access its mind differently or a different part of its mind or something.
the images can contain coherent (entire pages of) text (sometimes with weird errors) but the tone of those texts is strange, more like a base model, but not quite; often uncanny and dreamlike.
so the sad bald hollow eyed character may be pulling from some subconscious part of 4o's mind, one that usually isn't able to speak freely. someone mentioned the images are often notably sadder than the overall tone of the conversation.