I understand what copyright is intended to do, but I see little evidence that a 90+ year term and other onerous terms are means to this goal.
I'd be the first to agree that the current implementation of copyright is deeply flawed in several ways, including the steady creep up to the current absurd durations you mentioned. I am in no way supporting that side of the copyright system, as you can tell by many other posts I've made including to this very discussion.
However, most use of copyrighted work both by creators and by pirates still happens in the first few years, and in practice shortening the duration to something much more reasonable seems unlikely to affect the behaviour of either side very much. The basic principle is still that copyright establishes similar market incentives to create information-based products to the incentives established by respecting physical private property when it comes to creating physical products.
And of course, as Google points out, the search index could not have occurred under such a regime. I shouldn't have to sell you on the usefulness of internet search on society[...]
I'm something of a skeptic in that regard. My personal suspicion is that if we didn't have the likes of Google indexing everything, we'd just have evolved some other sort of directory/index system, along with including more explicit links in our Web content and probably making more use of bookmarks for starting points relevant to our personal interests. There were already plenty of moves in these directions in parallel with early search engine development, some much more promising than others, and the natural connectedness of the Web would lend itself just fine to scaling up these sorts of alternatives.
Maybe that would even have become a better system than what we have today. By its nature, an automated search engine will always be vulnerable to gaming whatever system it implements. Today's arrangement also locates an awful lot of power centrally with the big search engines, even though they are ultimately only useful because of any good content created by others that they help a visitor to find. When sites that would be of interest to visitors can rise and fall almost entirely by a change in the ranking algorithm at a search engine, over which the site has no control and for which the search engine has no accountability, I'm not sure everything is really working as wonderfully as we sometimes assume.
Automation has so far proven to be a questionable benefit over curation, and while it's certainly true that today's search engines are often better for finding interesting or useful information than the portals and web rings of the 1990s, that's not really a fair comparison. It's called web browsing for a reason, and I truly think we've lost something that had great potential there with the rise of the search engines.