AI which responds to players’ intentions is more fun, study reveals

Researchers at NC State University found that AI which is ‘intentional’ — and gauges a player’s intentions and what they are likely to want — is more enjoyable to interact with than AI which does not.

A study of 224 players conducted by researchers at NC State University found that AI which can infer what players likely want from the AI when they share information is more fun than AI which fails to do this.

They call this concept ‘intentional AI’. This kind of AI gauges how players are likely to respond to new information and what players expect from the AI. They designed a study in which humans played the card game Hanabi with AI with various degrees of intentionality — from no intentionality at all, to full intentionality. They found that the ‘full intentional’ AI was more fun to play with.

We think it’s important for AI programs to be able to communicate effectively with users. Being able to gauge people’s intentions is a critical part of how people communicate with each other, and we think our work here is a step toward incorporating that kind of intentionality into AI in a meaningful way

— Dr Chris Martens, NC State University

This points to more exciting future work around intentional AI programs. You can read more about the study, and their plans, here:

https://news.ncsu.edu/2017/12/making-ai-intentional-2017/

As a side note — we believe in this ethos of ‘intentional AI’ too here at Spirit, of course. Our Character Engine is built around meaningfully responding to player intention; and we apply this specifically through characters who can respond with personality and with their own narrative agenda.