Does machine-assisted production alter grammatical structures?

This paper asks whether machine-assisted production during informal writing changes the grammatical structures used by individuals. To simulate different production environments, we elicit essay-length samples from 166 American English speakers under six conditions representing different levels of machine-assistance. This data shows that machine-assistance is responsible for a majority of the words in each sample. But does this machine-assistance change the inventory of grammatical structures that are used? Our analysis shows that the syntactic similarities between samples remains quite high across conditions, tentatively suggesting that machine-assistance does not have deeper grammatical impacts.