There are multiple different strands of history in the term "RPG". To grossly oversimplify, western RPGs descended from Dungeon Keeper, Japanese RPGs descended from dating games :) Then the streams crossed, and western RPGs adopted JRPG tropes.

Visual novels (such as dating games) are a huge genre of game in Japan and the Far East, and have been for decades. In a visual novel, one very common convention is to depict the characters in the scene taking up most of the screen, and then a text box at the bottom containing what the characters are saying or thinking, with an attached portrait of the character currently speaking. Example:

Japanese RPGs were establishing their conventions at about the same time that dating games were, and so JRPGs' interfaces were informed by visual novels. Often JRPGs would include character portraits to help you visualise who you were talking to, rather than having to go by the few pixels that formed the character's face. Example:

I believe this principle is what carried on to the full-motion animated character portraits you're seeing in modern RPGs.

(Example screenshots chosen to illustrate the design principle, rather than for historical accuracy; the trope much predates both these examples.)

There are multiple different strands of history in the term "RPG". To grossly oversimplify, western RPGs descended from Dungeon Master, Japanese RPGs descended from dating games :) Then the streams crossed, and western RPGs adopted JRPG tropes.

Visual novels (such as dating games) are a huge genre of game in Japan and the Far East, and have been for decades. In a visual novel, one very common convention is to depict the characters in the scene taking up most of the screen, and then a text box at the bottom containing what the characters are saying or thinking, with an attached portrait of the character currently speaking. Example:

Japanese RPGs were establishing their conventions at about the same time that dating games were, and so JRPGs' interfaces were informed by visual novels. Often JRPGs would include character portraits to help you visualise who you were talking to, rather than having to go by the few pixels that formed the character's face. Example:

I believe this principle is what carried on to the full-motion animated character portraits you're seeing in modern RPGs.

(Example screenshots chosen to illustrate the design principle, rather than for historical accuracy; the trope much predates both these examples.)