Despite the well-publicized privacy headaches that this type of feature has caused other companies, Scribd launched Readcast on an opt-out basis. That is, the first time people used Scribd after the Readcast launched, the company showed them a screen allowing them to opt out of posting the document to their public Scribd profiles. If users exited the screen without changing the default settings, the documents they downloaded were added to their public profiles.
The move drew relatively little attention until last week, when Santa Clara University law professor Eric Goldman wrote about it on his blog. "I don't feel I got adequate notice of the change. Haven't they learned anything from Facebook's Beacon fiasco or Google's Buzz fiasco?" he wrote. "Since apparently the message isn't sinking in, let me spell out what should have been obvious: People don't want to automatically publicly announce the documents they are reading."
By today, Scribd revised the feature to make it opt-in. People now must affirmatively instruct the company to broadcast information about what documents they are reading to other users.
Also today, Scribd joined Facebook's controversial instant personalization program. But, in a significant departure from some other companies to use instant personalization, Scribd will only enroll new users in the platform.
The company says that when new users sign up for an account, they will be given the option to participate in Facebook's instant personalization platform; if users don't make a choice, they will be opted-in by default. In that case, the documents that users download won't show up on their Facebook profiles unless they specifically direct Scribd to share that information.
Scribd will, however, use people's Facebook information to recommend documents for them to read; those recommendations will not be broadcast to other Facebook or Scribd users.
Facebook's decision to roll out instant personalization on an opt-out basis and tell outside companies the names of their users without those people's explicit permission remains extremely problematic. But at least Scribd appears to be reversing course on a troubling decision to share data about what people are reading without their affirmative consent.
I'm not saying Scribd is or isn't guilty of privacy faux pas here, but I completely disagree with Goldman's comment about what people 'obviously' want.
Arguably, plenty of people want to broadcast the documents they are reading - and still more of us find it intriguing to be able to 'follow' what someone else has read.