We have AI built into some tools I believe, but I have never been told I had to use them. The truth is they don't work all the time for every situation and the client is more worried about user data accidentally getting scooped up and spending time warning us to never enter any users information anywere, even so much as notating a user saying they have a limitation that explains why we performed a task in a non standard fashion is a complete not happening.
So if someone said, "I am vision impaired," someone reading our notes would probably be wondering... Why the f didn't they just do a,b,c it would have been much easier. But they are worried if those notes get integrated into something the AI gobbles up in the future, they don't want to get sued for that user information to somehow be linked to them. As that could be considered medical data I guess.
The funny part is, if an AI does use that data for learning now, it may start trying to instruct or perform tasks based off of highly inefficient solutions designed to assist a specific disability
Yeah not to mention some of the data for searches for porn coming from states with age age verification checks are using VPN's to other states. (Like 80% of porn hub traffic from Louisiana dropped after the age verification laws, and VPNs went up across the board, as in those users likely were not just showing up as users from other states)