This is easily 3-4 years away still, if not more. At the time I was bored and frustrated. Bored that games were only full of violence with nothing to say. Frustrated that I thought I'd found a way to bring the point and click story ethos out in a new way and no one I pitched it to could see why it was worth achieving.
So what's the best way to draw people's attention? I know, make up a terrible pun in French that I probably misspelled. Yep, I was so aspirationally high-brow it hurt. Don't worry, I got better.
The main thought was to break down the problems in creating a truly gripping dynamic story into key pieces, centred on a bottom-up approach where stories emerge from NPCs with enough interesting & conflicting drives - and then to tweak those drives on the fly with a high level 'Author' AI.
How did I do? Well, Natural Language and Text-to-Speech are getting pretty good, if by pretty good you mean infinitely better than 2005 but still completely terrible. See Google Home fail with my daughter.
Back then I was pretty annoyed that the big leaps forward in games at the time were nothing but eye-candy; with the biggest newcomer being realtime physics. I wanted to have a dig at it. I wanted to blow it up in frustration that it wasn't really pushing gameplay forward (except in niche puzzle games). I was trapped in a world where a game was all about the game mechanics and everything else be damned. My assault - a barely noticeable sarcasm about the latest advances.
THERE'S HOPE YET!
Two LONDON, UK studios look set to tackle this.
BOSSA STUDIOS are looking to use machine learning to jumpstart their stories and draw on the talents of ex Valve hero Chet Faliszek who helped pioneer the Left4Dead AIDirector. Exciting.
INTERIOR NIGHT are made up from a bunch of ex Sony and Quantic Dream people and want to make stories for real people. More power to them.