Bot or human? Creating an invisible Turing test for the internet

I'm not sure reCAPTCHA is really trying to detect automated vs human interaction with a browser. The primary use-case is to detect abusive use. The distinction here is if I automate my own browser to do things for me on sites using my personal account may not be a problem for site owners, while a spam operation or reselling operation which generates thousands of false accounts using automation is a big problem that they'd want to be able to block. I think reCAPTCHA is tailored towards the latter, and for it not to block the former might be more of a feature than a bug.

LinkedIn, for example, doesn't care if you as a human are manually looking at all your connections one-by-one or if you have automated a bot to do it: it will lock you out the same either way.