Skip to content

Thinking about Images in Social Networks

Images are great, we love using them to communicate, and now its said that facebook posts now contain an image 50% of the time.

But with social networks that actually seek your data in order to sell it on…which is the majority, you have to ask yourself, how much data did you just give over by posting a pic?  They say an image is worth a thousand words…..but with todays imaging technology perhaps that should be 10,000 words!

Let me expand a little on how this works with todays technology. You probably know information like date and time is embedded in your digital photograph, but in fact there is a load more data than just that. The data is stored in the image itself in a format called EXIF. What you may not be aware of is that this data can also include your GPS information.  In other words not only what time you took the picture, but exactly where you took it!  To get a better idea download one of the many EXIF viewers from your mobile app store and take a look at the EXIF data of a picture you took.  Then turn on GPS and take another pic and see the additional info embedded in the picture.

The next issue is image recognition.  Facebook caused a storm of protest when they used facial recognition in their social network, and turned it on by default.  So if you happened to be in the background of someone else pic, you could have been identified.  Combine that with the EXIF data and you start to see how intrusive this technology can be.  Now technology has moved on to try and identify objects and brands in your pics….so social networks can sell this information on.  Combine that with associated text like “Little Jimmy’s Birthday”, plus the name of the person posting and you can see how quickly highly intrusive perspectives on your private life that you did not realise you were sharing can get shared.

So next time you post a pic to a social network…THINK first.

In the Krowd, we explicitly go in the opposite direction of other social networks.  We seek to minimise the amount of data we hold on our users, and also protect them from inadvertent exposure of information.  So how do we do this for images?

Well first we strip all EXIF data from any image that is posted and stored.  So only the date/time of your post indicates a possible date/time of the picture.

Second, we reduce the image resolution down to the minimum needed to view the picture on your phone or tablet.   This is important because most cameras take very high resolution pictures today, far higher resolution than the eye can behold (the idea is that if you blew up the pic to wall sized it would still look good), in reality you only need 3M pixels to be able to print a decent quality pic on A4/US letter sized paper.  Who needs more than that?  Well the answer is going to be the image recognition guys.  Because although the eye cannot perceive it there is image data stored that can see the letters on your clothing label, and if they can only see part of it, perhaps blurred, well they can apply digital enhancement technology to determine what is written.  More than that they are developing object recognition, to be able to tell one brand of shoe from another, or one chair make from another.

So when I say a picture tells a 10,000 word story perhaps it should be  100,000 word mini novel!

Be careful what you post, what you see is not all you are sending when posting todays high res images.

Tracking Location Creeps Us Out

Using Co-Location not Location-based capabilities to deliver a social networking service.

Lets face it, very few people turn on their GPS except for a specific purpose, such as getting map directions.  We intuitively understand that when GPS is on we are being tracked by the company behind the app and/or perhaps the mobile phone maker or network provider.  Its one thing for someone to know what information about ourselves we explicitly put in the cloud, its a whole other thing for some commercial entity to be tracking where you are and when you are there (Governments reserve special privileges in this regard…nothing we can do about that in our app).  What may be less intuitively obvious is that these same companies can correlate your location-based information with other location information sources (such as someone else’s mobile GPS information) so that they can look for patterns over time, inferring who you are friends with, predicting where you will be in the future. It may be cool to turn up at your favourite coffee shop and have your regular order waiting without even asking – but imagine that movement pattern information be hacked and used for more nefarious purposes?  It makes you shudder.  Imagine less scrupulous businesses using that information and what they could do…all within the current internet and privacy laws, or because you checked their long-winded T’s & C’s without reading them closely – who does?

They’ll claim all this data is anonymised to protect your privacy – but at what point does a specific movement pattern plus other anonymised data such as age, gender, work location, the fact you own two dogs, or a red ford car etc start to fully define you…just without a name attached?

We pondered the dilemma and challenge of delivering location-based mobile app value without know where you are – could it be done?

The answer was to stop thinking about location-based services, and instead think about Co-Location based services.  For a social network like the Krowd there is value in knowing who is nearby, but for the purpose of connecting co-located people together it became obvious we did not need to know where in the world you are, just that you are in proximity of others, and to do this we don’t need GPS turned on.

We go one step further – because of course we do know who you were recently co-located with – and that is information you may not want tracked over time. Hence we only store the last few Krowds that you were in – we plan some cool features around this data – we don’t think our customers will care if we know where they have been the last few days, but not the last week, month, year etc – if you do…leave a comment below. We’ll be enabling the tracking of this policy by our board of trustees who ensure the privacy of your data on our customers behalf, and ensure app feature changes remain in alignment with the policy of not storing information about our users except as needed to deliver the service they demand of us.

What is a Krowd?

In short its the social networking app created by Krowdthink.  But thats does not tell you much, especially as we have not launched the app itself yet, it enters 1st public test in Q1 2013, email here if you are interested in supporting any of the test phases.

Our start point for the Krowd concept was the crowd, a group of people congregating in a location because they have a common interest or purpose.  So by definition some of those people are people you may want to connect with.

But as a concept, what is a Krowd? We define the Krowd on our website as a trusted social network to connect like minds.

There are three key components to this statement: ‘trusted’, ‘connect’ and ‘like minds’.  Lets expand on each of these:

‘Trusted’ – we outline what we mean on our website and a previous blog,  here and here. These values are especially important when put into the context of connecting people.

‘Connect’ – when you combine online social networking and mobile devices you bring in the potential for location-based value.  We believe the ultimate value of a social network should be to facilitate the face-to-face meeting, yet no online social networks to date have made this their top service focus to their users. When we say connect, we don’t just mean virtually in the cloud, but potentially in reality face-to-face. So the Krowd app is as much an intellectual introductory service as it is a social network.  But to be able to make a decision to meet someone you know is nearby, you’d like to check them out anonymously, to validate they have common interests.  But we don’t want to expose our personal profile to everyone just because they happen to be in the same place at the same time. We need to be able to define our profile so that is contextual to the type of crowd we are in.  If at a sport event, we might be happy for people to know which team we support, how long we’ve been a supporter, our thoughts on previous matches we’ve watched or on particular players we’ve seen. But we may not want to expose other aspects of our social life. For example you might not want to expose the fact that you are deeply into fashion when at the sports event….or maybe you do…who knows, whatever, it should be your choice as to how many of the personal onion layers of self that you indicate to those in a crowd by means of introduction, and even once connected you still may only want that connection to be contextual to your common interest.

‘Like Minds’ – how do we assess someone is of a like mind before we meet?  at a sports event we maybe able to see what colours they wear, but online we don’t have the same visual cues.  So in the Krowd we enable people to be able to create a profile that is specific to the type of Krowd they are in.  The profile is based on a series of mini-personal-blogs we call Klogs (Krowd logs), that you build up over time.  People can comment on your Klog and you can respond to those comments.  When you do, that event is fed into the Krowd feed which combines real-time chat at the event with status information such as people joining or leaving the Krowd or posting or commenting on a Klog.  Its a real-time broadcast channel that you can use to chat with everyone in the Krowd.  If you want to get personal, invite them to a Bubble for a private, yet still anonymous, chat.  Meet face-to-face once you get comfortable or just engage with the virtual, but totally localised, Krowd conversation.

For those interested, I can point you towards many useful papers on the need for multiple identities/persona online, as a means of protecting identity or as a means of sustaining privacy in contexts you care about.  But here is my over-arching observation – current leading social networks are structured to be essentially one-dimensional in terms of the aspect of your persona you can decide to expose.  Linkedin for business, facebook for friendship/social (despite their attempts to be all things to all people, they don’t engender user trust to allow these initiatives to be properly embraced).  Google+ is trying to address this issue, but they undermined their own efforts with the demands for real identity.  Again its that trust issue that creeps in and creeps people out.  My first blog post was about Online Privacy – but in reality its a trust issue, and gaining user trust in your company and social networking app is the critical building block of a next generation social network.  We have mentioned location based capability in this post – and nothing creeps people out more then the idea that where they are and when they are there is tracked….in my next post we’ll discuss our thoughts ideas and mission with respect to achieving a location based service that has no check-in needs and does not know where you are – and does not need or even want to know.

What is Online Privacy?

At Krowdthink we seek to offer our users online privacy as a foundational building block of our service.  So we had better be clear from the outset what it is, or at least to offer up our definition. Even the wiki definition of internet privacy struggles to express this concept succinctly, but we will try.

One thing is clear from reading a plethora of learned writings on the subject – privacy is user defined.  We all have our own standards as to what we consider private and how we wish our privacy to be respected.  Everything from the standards our parents instilled into us, the painful object lessons learned in life, to the culture of the country we live in, through the religious and other beliefs we hold dear, all these define what we consider to be private.  Yet at the same time, we are social animals, continuously accepting that through social interaction we give up some small element of privacy.  Only a hermit could properly be defined as a private individual.

The Social Network is now woven into the fabric of our society as delineated by the leadership position facebook has taken, but Zuckerberg, while oblique about his position on privacy, is clear on issues of transparency and identity. See this awesome Venturebeat write up – and the associated commentary. So if we are to use social networks and maintain online privacy, the question is how?

Our answer is to think in reverse – rather than what is online privacy, for which there is a unique answer for every person in the world, instead ask the question ‘when is my online privacy intruded upon?’  For which we believe there is one answer:

“When the information I provide is used for a purpose other than that for which it was understood to be provided”

There are two requirements for addressing this – the first is that the service provider for your social network must make clear precisely how your provided information will be used, and this definition should be simple, explicit and readily understood.

The second is that the users of the social network service should be encouraged to act as good privacy citizens, through online social network policy and tools, as well as mutual respect.  It is notable that as soon as facebook opens up a new feature and makes your information more public, its actual usage levels go up – we humans are not, by nature, nice – we are happy to intrude on others privacy but less than willing to accept it when our privacy is intruded upon.  Fostering positive privacy etiquette is the responsibility of the social network provider.