Romantic Ai Philosophy & Electra Progress

Romantic Ai Philosophy & Electra Progress

The newest version of Electra represents a very mindful reflection on the graphics and flow of using the writing tool. Pairing the 3D flow chart with the text message output remains the bedrock of the app. One of the key missing features was clear connection between tone properties of the player options and the impact on the ai character’s reaction and chosen response.

Now that symbols have been developed in game to represent these factors as the player plays the game, they can be placed next to the corresponding tone sliders in Electra. This allows writers to focus on options they want to provide to the player without checking the calculator tab for impact.

I have also now exposed the filter metric for processing player tone choice and Ai response. The default processing formula is Tone (ai character personality’s positive and negative reaction to each tone type). But in many contexts, such as romantic, Charm can be selected instead. Charm is the attraction/dating counterpart to the Tone metric. It processes the player choice properties differently, putting priority on what the ai character’s personality is attracted to or not rather than what they platonically like or dislike. Something fantastically interesting about this approach is that some characters will actually be attracted to qualities that they otherwise very much dislike.

This ambiguity allows for a very intentional strategy for players in which they can prioritize the Friendship (Bond), or the Ai’s Attraction to them (Spark). Of course players can also choose to ignore the questionable morality of tactics for emotional dialog option selection and choose based on purely what they want to say and how they want to say it.

Ideally, given the system reflects real life successfully, player’s choosing to express themselves naturally and in line with their personality will cause a self selection of developing relationships with Ai’s that are drawn to who they are and distancing Ai’s with hostile or non-receptive reactions to them.

Of course the creator’s writing style also plays a huge role. Also similar to real life (assuming natural player expression), there could well be Ai’s that are absolutely in love with the player (like their choice in expression, and are attracted to how they express themselves as well), but the Player may not be interested.

I think that creates a really realistic dating environment where the combination of all relationship dynamics are possible. Because of the way the system is setup, Player’s options are each centered around the Ai they are talking to. In some ways this is challenging because it can easily cause an inconsistency in available options and types of expression. For example Funny & Mean options (teasing) may be relatively light hearted and enjoyable talking to one girl, but another girl’s conversational options that are property wise still Funny+Mean may feel overly cruel or tasteless.

So in a fascinating way, the player is not just selecting for Ai characters that accept them for who they are, and are attracted to them, BUT also choosing between Ai characters based on who they themselves (player) are allowed to be when they talk to that character.

Metaphorically this is also similar to how in real life, you are significantly impacted by the people you keep as friends. They encourage or discourage you to be a certain way.

In a literal sense however the system is limited in that there is not a direct user input to the character beyond expression choice (emotional options), physical actions (eye contact, hugging, etc), and the accumulation of other character interactions and relationships that the Ai is aware of.

In other words, the player can not type out a completely original response to the Ai. But I absolutely believe and can make a strong case that original input with limited meaning in response (traditional Ai approach) is inferior to limited input options with full meaningfulness in response.

Is the expansion of the tool to include new timeline parameter, variety in exchange, multiple metrics, and multiple character conversations causing an exponentially intimidating effect for writers/users of Electra? Time will tell. I do think that Electra has a high end skill cap that is almost limitless.

It requires the skill of conversational writing (both sides of a conversation) exponentially expanded based on the Set of All of meaningful options and sub-expression choices.

That said, the tool was initially created purely for my own use. And so almost all the features and design were inspired by my own desire for the best possible experience and efficiency in writing advanced ai characters into the game and hopefully other games one day.

So I find it useful for many objective reasons to have all these new features and graphics so I expect other users will as well. Unlike video games which must contemplate access to casual users, creation tools should significantly consider peak performance design decisions and systems as higher priority than a easier learning curve.

Especially given the system being potentially one of the most difficult content creation tools ever created, I have to take care of the people that I know for sure CAN use it. This goes out to all the pros in the process of learning it already and the people who will pick it up one day: I hope you like what I’ve done with the place!

I fully expect to be making changes and updating it as I go however so feel free to reach out and offer suggestions, criticism, or encourage expansion of existing systems. I am also down to make the case for the superiority of all the core principles and design decisions regarding this approach to Ai in both platonic relationship building, as well as the romantic dimension.

I am heavily engaged with philosophical research and invite all theoretical objections, challenges or curiosity. And congratulations in becoming a pioneer of 3Dimensional writing, should you pick it up or continue to use it!


Leave a Reply

Your email address will not be published. Required fields are marked *