The deal would allow Google to digitize and sell books at prices set by a new book registry, a collective rights group similar to the music industry's ASCAP and BMI. Civil liberties organizations have pointed out that the agreement leaves Google in a position to amass at least as much in-depth information about users' reading habits as libraries.
For that reason, groups like the Electronic Frontier Foundation have said the settlement should have terms obligating Google to protect users' privacy -- such as provisions requiring the deletion of loggin information. Instead, the amended pact merely says that Google won't share private information with the registry without "valid legal process."
This promise doesn't go nearly far enough to solve the privacy problems posed by a digital book registry. First, requiring "valid legal process" doesn't set the bar all that high considering that any judge can rubber-stamp a subpoena requiring Google to disclose information about readers. Sure, Google can challenge subpoenas in court, but nothing in this agreement appears to require the company to do so.
In other situations, Google has notified users about subpoenas, which at least allows people the opportunity to argue that subpeonas should be quashed. But individuals aren't always in a position to hire lawyers for that purpose.
Of course, Google might be planning to hold itself to higher standards -- including notifying users about subpoenas and giving them the opportunity to object before the company complies. Google might also be planning to fight the subpoenas on users' behalf. But it's not clear that this agreement requires the company to do so.
The reality is, as long as Google plans to collect and retain information tying users -- or even IP addresses -- to reading material, users' privacy is vulnerable. The EFF and ACLU are warning that without more built-in protections, Google could "become a one-stop shop for government surveillance into the reading habits of millions of Americans."
If Google doesn't want that to happen, the company should agree to some new limits on its ability to collect and retain data about the books that people read online.