Analysis of Language Change in Collaborative Instruction Following
This addresses how language evolves in collaborative AI-human systems, showing a counterintuitive trend compared to simpler reference games.
The study analyzed language change in collaborative instruction tasks, finding that unlike prior work on reference games where language complexity decreases, instructors increased language complexity to better collaborate with increasingly skilled followers.
We analyze language change over time in a collaborative, goal-oriented instructional task, where utility-maximizing participants form conventions and increase their expertise. Prior work studied such scenarios mostly in the context of reference games, and consistently found that language complexity is reduced along multiple dimensions, such as utterance length, as conventions are formed. In contrast, we find that, given the ability to increase instruction utility, instructors increase language complexity along these previously studied dimensions to better collaborate with increasingly skilled instruction followers.