Hacking Bullies
Hacking Bullies
Saturday, 13 March 2010 09:55 Hrs
❝ It has now been several months since I presented the talk at Trampoline. The idea, a simple hypothesis, 'is there a better way to identify and disrupt Cyber Bullies?'

Several months later

This is an edited talk I presented at Trampoline 2 [0] in Melbourne [1], Australia, Saturday 24th of October, 2009 entitled: ‟Hacking People, Hacking Bullies”.

❝ Is there a better way to identify and disrupt Cyber Bullies?

I know for example the current best detection and prevention systems we have, the Olweus Bullying Prevention Program (OBPP) [2] by Dan Olweus [3], manually identify bullies by peer association in their social networks. [4] Could this technique be extended to on-line social networks? I wanted to explore the idea of using network science techniques using maths and computers to identify hidden bully clusters in social networks.

The talk was divided into three parts: a) ‟Know your Enemy, Hacking humans”; b) ‟In the field: real life examples” and c) ‟Hacking Bullies”. [5] I tend to write about these ideas to look for new insights into problems I see. [6] Bullying is a problem. I’ve experienced it first hand. It is the prime motivation for exploring these ideas. Bullying on-line is an even bigger problem and in an age of ‟social software” worth exploring further.

Preamble

My name is Peter. I’m a programmer. One of those gen-X slackers you hear about. My second computer was an Apple 2e clone shipped from Singapore in parts and hand assembled. I’ll be as lo-tech as you can get and read from sheets of paper. Occasionally I’ll digress from the script. I tried to Kevin Bacon to stand in but he was busy and suggested Derek Zoolander and as soon as I mentioned to Derek there would be lots of big words and no cat-walk - Derek bailed, so you’re stuck with me.

❝ My talk is on bullies and what happens when they discover computers.

If you want to use twitter and hash tags try #hackbully [7] I’ll probably read the comments later. I’m not a psychologist, I’m a technologist with crap social skills. I’m interested in how bullies are adapting to computers and network, how to identify and neutralise them.

Hacking Bullies

The Internet is a big network. Bullies now have a much bigger playground and audience to choose from. So lets look at some real world examples. Some people collect stamps, others prefer spore, moulds and fungus; mine is on-line behaviour.

Here’s a few examples of cyber—bullying I’ve found. Gretel was on live national television and Will Anderson (ATTACKER) decided that Gretel Kileen was good TARGET and posted live insults on twitter taunting OUTSIDERS to join in on the fun. Gordon Ramsey, a known ATTACKER gets lots of attention these day when OUTSIDERs look for images of ‟fat Gordon Ramsey”, for fun. Is this a case of OUTSIDERs fighting back? Erin Andrews, a news sports reporter in the US was ATTACKED by having video footage of her walking around after coming out of a shower posted on Youtube. Then had trouble convincing police she was TARGET.

Bullies have discovered computers and networks. Even Mika uses the power of song writing to asymmetrically attack bullies from his childhood. I think it’s time to use our powers for good. What can we do?

Network science 101

Now we are almost up to the fun bit, where we use out knowledge of network science to hack bullies. But first a short diversion into Network science theory. Understanding networks allows us to understand markets, societies, species, corporations. Networks can be biological man made or social.

If we understand the rules of networks it give us the theory to create tools to hack bullies. There are four ideas that make hacking bullies possible: Small world networks, hubs, how information spreads and mapping networks.

1) Small world networks

Steven Strogatz and Duncan Watts are both mathematicians who succeeded in understanding the properties and structures of small worlds. Strogatz/Watts started asking basic questions about connections and relatedness of nodes in networks and realised understanding the ‟relatedness” of nodes and how each node is connected to another node is an important problem.

Strogatz/Watts, first tested their ideas on a network model of Hollywood actors and their relationship to ‟Kevin Bacon#8221;. They had a theory but no confirmation. The problem to solve? If an actor plays a part in a film, what is the chain of actors to reach Kevin Bacon. How many links is there in the chain? Lets try an example. How is Derek Zoolander (Ben Stiller) related to Kevin Bacon?

❝ ‟Ben Stiller” was in ‟Tenacious D in the Pick of Destiny” (2006) with ‟John C. Reilly” who was in ‟The Wild River” (1994) with ‟Kevin Bacon”

Strogatz/Watts discovered two structural properties of what we now call *‟small world systems”. The first, there are relatively few related nodes in a small world. The second, small world systems tend to cluster together closely. The cluster close enough to find another node in a distant small world via minimal hops.

2) Hubs

❝ Knock a hub out, the network is severely interrupted

Albert-Laszlo Barabasi looked over the results of Watts/Strogartz data and made an interesting discovery, networks are not connected randomly. Barabasi found the mechanism that lets small world link together. The idea he observed is that some nodes in a network have many more links to them than other nodes. We call these super nodes, hubs. The two most important insights Barabasi has made: a) Networks are robust. If individual nodes are knocked out, the network survives. b) The hub is the weak spot. Knock a hub out, the network is severely interrupted.

3) Spread of information

Alessandro Vespignani is a physicist who specialised in understanding how disease spreads in biological systems. By reading Barabasi work on networks, Vespignani observed transmission of disease in a network is not random. Diseases moves quickly and spreads far via hubs. Conclusions: Small changes in the network have an effect on the whole system because the network is inter-related. But a virus, disease or information will not be eradicated in a network unless you take into account hubs. Understanding Hubs is the key.

4) Marc Vidal and genetic maps

Marc Vidal is a geneticist interested in looking at the inter-connectedness of cells. After reading the research by Barabasi decided to look at disease as a network. Instead of looking at the connected cells one by one Barabasi suggested to Vidal to look at the connections of disease and genes to create a map. A map of related diseases and genes. This idea is significant. It means you can now see how diseases are related. An idea made possible because it was redefined as a network science problem.

Now we come to our ultimate aim. Can we hack bullies who inhabit our electronic networks? Can we use network science for example in the same way terrorist cells are identified and disrupted?

Predicting behaviour

Could we use network science to predict where bullies might be operating? If so, how might we go about this? Lets think about what we know about bullies.

  • bullies inhabit small world social networks online and offline

  • bullies have identifiable negative behavioural characteristics

  • bullies attract followers who attack,defend or remain neutral

  • bullies exhibit hub like behaviour

  • bullies have victims

What would happen if on a social network service we looked at reported behaviour and people at the same time. Just like Marc Vidal and genetic maps. What if we mapped the behaviour of social networks to look for related behaviour and hubs? What we would be looking for is the hubs of bad behaviour. Could we identify bullies by their associations to HENCHMEN, OUTSIDERs who observe an ATTACK? We know from Barabasi that if we ignore the hubs the problem remains.

A network of bad behaviour is stable because the hub remains intact. Can we knock out hubs and interrupt bad behaviour? There is nothing to stop any of the big social software site doing this. We now have enough theory to create the tools to disrupt bullies in networks. We know from Vespignani that small changes in the network make a difference and that if we redefine the problem and think of bullying and bad behaviour as a network science problem, we can tackle it successfully with maths instead of ignoring it?

Reference

[0] Trampoline 2, ‟my images of Trampoline 2 taken on the day on flickr”, [Last Accessed: Saturday 25th April, 2015],

  • http://flickr.com/photos/bootload/collections/72157622546767967

[1] Melbourne, ‟my flickr set of Melbourne”, [Last Accessed: Saturday 25th April, 2015]

  • http://flickr.com/photos/bootload/sets/72157608350016296

[2] olweus.org, ‟Olweus Bullying Prevention Program”, [Last Accessed: Saturday 25th April, 2015],

[3] Dan Olweus, #8223;History of Dan Olweus”, [Last Accessed: Saturday 25th April, 2015]

  • http://www.clemson.edu/olweus/history.htm

[4] Olweus Bullying Prevention Program, ‟The Olweus Bullying Prevention Program Overview: (Research_OBPP_Effective.pdf) Black, S. A., & Jackson, E. (2007). Using bullying incident density to evaluate the Olweus Bullying Prevention Programme. School Psychology International, 28, 623-638.”, [Last Accessed: Saturday 25th April, 2015]

  • http://www.olweus.org/public/document/obpp_effective_research

[5] Paul Graham, #8223;What you can’t say”, [Last Accessed: Friday 2th April, 2015],

  • http://paulgraham.com/say.html

[6] seldomlogical, ‟Don’t read this: Don’t think I write for anything, other than for my own selfish reasons. So stop reading, right now. Stop!”

  • http://seldomlogical.com/2009/AUG/27/don-t-read-this

[7] #hackbully, ‟a twitter hashtag for responses on the talk. I was interested to see what kinds of responses I would get on the back channel.”

  • http://twitter.com/#search?q=#hackbully
―~♞~―

My neek verification.

bio Another Scrappy Startup ☮ ♥ ♬ ⌨

short <seldomlogical.com/hackbully.html>

contact Peter Renshaw / 🐘 — 🎨

← HOME ↖ UP TOP ↑ BEST →