I don’t think it farfetched to suggest that you could be ‘fingerprinted’ as a user of even a public computer by patterns that have been detected in the way you use the computer. The web sites you visit, in what order, how often you comment on stories, how long they tend to be, how much you loiter on each site, your vocabulary… AIs are very good at picking up patterns and relationships we didn’t even know existed.
An insurance company could definitely fill out a detailed risk profile of you by looking at your internet usage. Won’t it be fun when we find out our car insurance or life insurance is being ‘adjusted’ based on our online profiles? Talk too much about your late-night cookie eating, or how much trouble you are having losing weight, and next year your life insurance goes up 20 bucks and you have no idea why. Or talk too much about hating government and spend a little time on gun sites, and suddenly you are getting a visit from the FBI.
You go for your next job interview, and it goes swimmingly. As a next step, the company orders up a report from ‘research analytics, inc.’, and they find out how many hours you’ve spent on the internet during business hours, whether you’ve bad-mouthed your previous employer online, whether your location data shows you’ve entered legal marijuana stores, how late you stay up at night, whether your online habits indicate you are politically undesirable, ad infinitem. You don’t get the job, and you’ll never kmow why.
Whether these things are being done today is unknown. Whether the data collection going on now enables that kind of behaviour is indisputable.