You claim “Google is modifying the content” but I think there is a significant difference between modifying/redistributing and someone modifying content using a program Google wrote. If modifying someone’s content, for personal use, with a program without the copyright holder’s permission IS wrong, then I think a whole class of browser accesibilty and development tools are then made illegal. Things like the Web Developer toolbar for Firefox, screen readers for the blind, and client-side stylesheets all now become suspect under that interpretation.

William: as far as I know, no copy of the page is ever “distributed”; the page isn’t sent to Google and no extra copies are made, so I find it hard to make an argument that the CC license is violated. You might have an argument that the copy resident in the user’s computer’s memory after AutoLinking is an infringing derivative work, but such an argument would make most, if not all, popular Firefox extensions illegal and would, at the very least, outlaw any browser which costs money or displays ads to the user (good-bye, Opera).

How is there any difference? From an end-user perception the effects are indistinguishable. A judge wouldn’t let Google hide in a technical loophole by claiming they modified the content on the user’s PC rather than on their servers. Either way you cut it, Google is modifying the content. If this action is deemed legal it will be because of something like “at the user’s request” rather than due to technical issues like where the bits are stored.

Let me point out, that Google may in fact be luring the end user into violating the terms of the CC. When users hit the AutoLink button, a copy of the content is “distributed” to google.com where it is analyzed. Did the user adhere to the CC? Was attribution given? Was it for non-commercial purposes? I could argue that sending content to Google so they can insert advertising (even if I’m the only reader) is a commercial purpose.