Thanks for the link. I agree odds would be slim, but this can potentially set a precedent for others to do similar things. Imagine, in the future, a new company comes up with similar add-in claiming to scan ads, and take more data than needed. Facebook needs to be careful of that too. (though i think that is not the reason they are shutting this down, just that privacy is favorable to them in this case :)). There has to be a better way to do this, and that needs serious discussion.

Also, as an aside, the add-in records the following things: country, age, gender, language etc. and the text of the ad. Not so clear what happens when ads have personal info (or friends' names) attached like xyz likes page. Payload seems to suggest they take every bit of info so it could include that as well maybe.

What is the alternative? No visibility for the general public into hyper targeted ad campaigns?

If the users are opting into this, isn't that the same argument for letting them opt into the Facebook platform itself? Why must we protect them from NYU and not from Facebook?

with respect, thats not the issue. With facebook they only collect "limited" information. Its what they do with it afterwards that matter.

How is it stored, is it secure, how do they anonymise it, who has access to the raw data, what happens when there is a break, is there monitoring, etc etc etc.

They are making a massive honey pot, and we just have to trust them that they are doing good job. (much like facebook....)

Please don't throw the baby out with the bathwater. There's no obligation to opt-in, and there's risk in everything we do and don't do.

The tool would need to guarantee that the data of a user opting in doesn't contain private information about his contacts which may not have done so.

Facebook doesn't have the greatest security track record. While it wouldn't be productive to argue that one group is inherently more trustworthy than the other it's worth pointing out (as others already have) that the code is available to the public.

Code being public doesn't mean its secure. See php, django, gnupgp, openssl etc etc etc.

Again, the issue isn't the collection, its the storage, processing and disposal that really matters here.

> its the storage, processing and disposal that really matters

That you can say about almost all web apps? Then almost maybe you shouldn't use the Internet, I start thinking

The last part is the crux of this dispute to me. If the users are opting into this, isn't that the same argument for letting them opt into the Facebook platform itself? Why must we protect them from NYU and not from Facebook?