Some quick links - Brave new web

Although I have not been out of office for so long (CCCamp and some other stuff) my dear office mates had the felling I am permanently on the road.

During the last time I collected some topics I would like to write about, which I will do today. Here are some links I just want to mention quickly.

Spock seems to be the new privacy killer. It harvests data about people from many different sources and users can add tags and even photos of others. This gives me a quite bad gut feeling. A prevention method that most social networks offer might be to make your profile only accessible by other members (well, or not to use any of them at all). A comment on netzpolitik.org (sorry, in German only) gives a hint how to sabotage this platform.

Scary, scary number two: Polar Rose offers face recognition via a browser plugin. It tells you who is in the photo you are looking at. Don't know if it works properly already but I don't like this.

An analysis on Blog Age (German again) shows that many web services make is hard for users to get rid of their account. I recently had a similar experience where I had to mail the support team to delete my account. That's not good practice!

To be no only pessimistic here are some good news: freebase is now in "public alpha" so everybody can read data while adding data is still only allowed to registered users.

Also nice: I can recommend the current CT Quarterly issue which has the title “The Coming Revolution in Scholarly Communication & Cyberinfrastructure”. [via science commons blog]

After some month of waiting: Wikiproteins is online since beginning of August.


Nikolaj Nyholm hat gesagt…


I'm not sure whether this will calm you, but we are very aware of the potential privacy issues.

The approach that we are taking is talking to various privacy groups about one of them maintaining a black-list of people who don't want to be programatically recognized. It won't be easy but we'll do our best effort to make it work on our side.

/n (who is envious that he couldn't make it to CCCamp)

Konrad Förstner hat gesagt…

Thanks for your comment, Nikolaj, but if people are not aware of this black-list, they will not use it. A white-list could be a solution: Only people who give explicitly the permission could be detected. But even this might not really work. I doubt your software is 100% precise. What if my face is linked to a wrong name, e.g. of somebody who just was in the press after committing a crime?

At the moment I think although the technology is cool, it is Pandora's box.

Nikolaj Nyholm hat gesagt…


A white-list would be much like asking every website to have a robots.txt file which explicitly allows spiders like Google to index.
I can already today get public, but anonymous, images of someone into Google's index simply by naming them in a hyperlink.

I think that we can't get around the point of education, here. Many people don't understand that public space has been made global due to the net. If I made a fool of myself in Copenhagen city hall square - arguably a public space - there were maybe 100 people who could see me.
Now, city hall square is global public space because someone snaps a photo of me and posts it on her blog with a full description to go.

Finally on the point accuracy, you are right that we cannot match reliably across the entire web today. That obviously also goes the other way and means that we can't detect all pictures of you from the one you have on your about page.


Konrad Förstner hat gesagt…

I thought a little bit more about it. Although the robot.txt way would be a technical solution the white list concept would not work, too. It would give only the choice to the people who put the content online, not the ones who are on the media. So the "victim" could only sue the person who puts the content online.

I agree, educating people regarding this is absolutely important. But in a way it is very sad. A army of private cameras is potentially watching every step of our life so we are living in the constant fear of saying/doing something that we might regret later (and the web does not forget due to archive.org etc.). This is not your fault. But Polar Rose will make the automated harvesting of this pool of media much easier. Yes, at the moment it is just a browser plugin, but it will not stop there.

Well, we will create any technology that we can create. We have also nuclear bombs and have to deal with that risk. I guess you won't stop Polar Rose just because of a blog posting and it would be just a question of time and somebody else would create something similar.

So if we cannot control the development of technology we have to make sure that the society is using the possibilities wisely. Unfortunately the current trend does not look promising especially regarding surveillance/privacy.