Just over half of respondents said that they regularly disable AI functionality in some or all security tooling due to a range of considerations