WiFi Shall Speak Peace unto Wifi?

Just to show that I can change my mind, I’m going to… well… change my mind.  In my last post, I said that I could not see any way of using Computer Mediated Communications for conflict resolution in an environment where social order had broken down to the extent that there was no reliable electricity infrastructure, let alone broadband Internet access. Now I see a way.

The answer to some of those problems is a device like the OLPC XO-1 the brainchild of the One Laptop Per Child (OLPC) project which aims to provide developing countries with… well… once laptop per child (today is ‘national stating the obvious day’ – bear with me).

First off, these machines are designed to require so little power they can be charged for hours from a simple human-powered generator – the original idea was a hand-crank operated by pulling in a string (rather like a miniature outboard motor). That takes care of electricity outages.

Secondly, OLPC machines have a WiFi set-up which can not only connect to the Interent  but also to other OLPC. This works rather like peer-to-peer file sharing networks.  If Henry’s computer is within range of  an internet connection,  he can share it with all of computers that are in range of him. They can share it with all of the computers that within the range of their WiFi  & so on. The last computer in the chain could be miles away from the Interent connection, yet still be able to use it.

The computers can also communicate with each other. Such a system is essentially anonymous, as no individual machine can be tracked to a fixed base. If machines are distributed across a divided community there is no way of knowing who is who.

Would this bring such a community together again? The individual people operating such a network would not know which side of the divide the people they were talking to stood on. They would be forced to judge others, in the words of Martin Luther King Jr  “Not by the colour of the skin, but by the content of their character”.

On the other hand, hate speech would be facilitated to – participants could fully express their bigotry & provoke new hatred with the assurance that their would be no direct consequences for them.

In any form of human communication we seem to have a choice between empathy & deception. The capability for one cannot be divorced from the capability for the other. In this regard, there is nothing magic about CMC –  human nature is still in the loop, after all.

Some form of moderation could be used – but the moderators would have to come from inside the community & the question arises “who guards the guardians”. Of course, moderators could intervene from outside – but that is condescending that the only unity it is likely to produce is a common hatred of the moderator.

The same social dynamic that operates in real life is likely to be reproduced in the virtual world. Giving people computers is not a panacea. Behind all communications is still good old human nature.

The social psychologist Philip Zimbrdo’s classic ‘Stanford Prison Experiment‘ concluded that much of what we percieve as inherent evil in human behavior is actually caused by situational factors. Perhaps this virtual world could establish itself outside the real life situation to such an extent that it would break down barriers.

Perhaps a common project with no sectarian implications would help. Perhaps the form of interaction could be designed to encourage cooperation & discourage hate speech without quashing diversity – e.g. an eBay style reputation-based system that puts low-rated contributors at the bottom of the new messages pile.

In conclusion, it may be possible to use CMC in conflict resolution, but there is a lifetime’s work in there. Might be a satisfying lifetime, though.

A Single Candle

A few days ago, I attended a presentation for Arrow a local organisation that aims to use art & theatre in conflict resolution. This turned out to be eerily relevant, when another spate of utterly pointless killings afflicted Northern Ireland soon after.

Human beings have a scary tendency to label groups of people as ‘other’. Once we have taken that step, we do have to mix with ‘them’ – we think we know all that we need to know from simple stereotypes. It’s a seductive mistake  because it makes life so much easier – all of the complexity & ambiguity that makes it so difficult melts away. The final step of this process is to completely dehumanise the ‘other’ – & once we have taken that terrible step, violence & killing are so much easier.

As I understand it, the aim of the Arrow project is to shatter this illusion by bringing individuals together in theatrical projects. In the process, those on opposite sides of a destructive cultural divide may see themselves anew – as real people, with common human feelings & values.

It was suggested that aid was more important than self-indulgent self-expression – that people in war zones need food & shelter, not a part in a play. Well, refugee camps provide food & shelter – & the conflict that people seek shelter from often breaks out, in microcosm, within them. Material aid treats the symptoms, not the disease.

Besides, it’s not an either/or choice – you can have both.

It might not work ata ll. It might hardly work – but isn’t it better to light a single candle than curse the darkness? I found myself in sympathy.

The question that was posed afterwards was how to use computer mediated communications devices to keep in touch with the local people who administer the project. Later perhaps, those channels could perhaps be usedd to pursue the organisation’s aims through those devices by allowing people on opposite sides of a cultural divide to communicate safely, & establish virtual social networks that may help to bump start real ones.

I have pondered this, & come to the conclusion that the greatest obstacle is infrastructure.

Even if you had the budget to give people in computers & mobile ‘phones & deliver training in their use through those devices, what’s the point when they don’t have a cell network, can access the ‘net only through cyber cafes, & cannot even rely on the electricity supply?

CMC has brought about considerable changes in our peaceful, complex & interdependent society. However, the infrastructures that support these networks are delicate. Before we feel smug, we should perhaps consider how well we could get on if all of these things were taken away from us tomorrow.

Any suggestions?

Is There a Person with 100% on eBay in the House?

Here in England, Her Majesty’s Government is still completely missing the point of the information age.  According to the BBC, Gordon Brown is considering creating an eBay-style feedback system for public services (including the NHS) which is supposed to “[…] make use of the enormous democratising power of information.”

This is wrong on so many levels.

In the cunning plan detailed above, patients will be commenting (among other things) on the performance of their heart surgeons. I don’t know about anyone else, but I don’t know enough about medicine to be able to do that, & even if I did I could hardly be expected to be objective. If feedback is to work, it must limit itself with an area that participants are capable of making informed judgments about.

Also, feedback needs to be focused. There is no point in every NHS patient commenting on every aspect of their experience – these experiences are so varied that it would be prohibitively difficult to glean any useful information from the data. If the NHS tries to bring some focus to the process, for example by limiting feedback to answering specific questions about particular areas, it is all too easy to ask leading questions & avoid issues that it does not want to discuss.

Finally, feedback needs to be be two-way – ideally, both parties should be able to rate the other. If you are going to allow patients to ‘rate’ surgeons, the surgeons should have the same power. A patients rating of “Sick again only months after my operation – bad job” might not look so scandalous counterbalanced with “told he will be dead in 6 months if he doesn’t stop smoking 40 a day. Post-op, he smokes 50 a day. Idiot”.

Of course, physicians are prevented from doing that by the strictures of  medical confidentiality (not to mention having better things to do). The ability to speak candidly is another essential requirement of a valid feedback system.

This proposal provides only the illusion of democratic involvement. It is also condescending – ” let’s let them think they have had a say, & they won’t act up so much – it’s not as if it gives them any real power”.Maybe it’s a good idea, after all – but only from the point of view of politicians & bureaucrats.

There is already an better feedback system in place – it’s called an election. For an election to work properly, voters need all of the relevant information. ‘Giving’ us an eBay feedback system instead, is just an attempt to distract our attention from the fact that such information is routinely withheld from us or distorted by PR.

Poor feedback does not require you to leave office. Gordon Brown is not going to have to resign if he scores less than 40% (which is unlikely in the first place, as he is setting this system up).

Feedback & reputation systems do work – but only when they confine themselves to very limited areas, preferably areas amenable to objective measurement – “did my purchase arrive on time, was it well packed, was it good value?” In discussion groups, reputations can be acquired by asking questions like, “Did this reply solve your problem?”, or “did you find this comment insightful?”

Such reputations are not transferable. If I am a member of a group that discusses the music of Bach (which I like) & have a good feedback-based reputation there, there would be little point in even mentioning this in a Hip-hop forum as the two areas have next-to-nothing in common.

As we have many reputations in many different arenas it might be a good  idea to develop a system that would independently collate them all on a single page. “Quick payer 90%, Good Seller 90%, Tellytubbies expert 100%, Hip-hop dunce 10%”.

Finally, the best reputation in the world is not necessarily valuable. The fact that you have 100% from 1,000 comments might sound impressive, until you learn it’s from “www.aliensaretryingtostealmybrains.com

Gordon Brown / e-democracy forum rating / 3% – must try harder.

iSocial Skills

Yesterday, I argued that social media were looked down upon by some as a poor substitute for real-life encounters. Today, I’d like to turn their argument on its head.

Perhaps the people that believe this have had a poor online experience not because users of social media have deficient real life social skills, but because they have deficient on-line social skills.

The social conventions that make communications technology work have traditionally lagged behind its installation. I am old enough to not only remember a time before mobile telephones, but also a time before affordable landlines. I was 19 before I had a ‘phone in the house, & distinctly remember not liking my first telephone conversation (in a call-box) because I couldn’t see the other persons face.

I got over my problem with the telephone. In the age of ubiquitous mobile ‘phones so, it seems, has everyone else. When people talk about impoverished communications channels, the ‘phone is rarely mentioned. I learned to make sense of ‘phone calls in the same way that person who becomes blind learns to deal with the world again. I learned to partially compensate for the loss of one sense by paying more acute attention to the others, & people who grew up with telephones culturally absorbed that adaptation from the previous generation.

This process has taken place even further back in the history of technology. Soon after the telegraph system was first established, operators learned to recognise each other from the way they used their Morse code keys. Each person has a recognisable ‘hand’.

The same thing seems to happen with the advent of every new technology. When email first went mainstream, there was much talk of  ‘Nettiquette’ ( Internet etiquette). This consisted of some common-sense rules about interacting through plain text. For example, we were advised to be careful about jokes – without the verbal & body language clues, a gag that would work face-to-face might appear insulting or offensive in an email. The smiley was born to obviate this risk.

Once again, people have become used to email. The conventions that developed organically are so taken-for-granted now that they have become effectively invisible – you only know what they are if you can remember the brief period before they put in an appearance.

Now the rate of change has accelerated – new computer mediated communications media appear frequently,  & enjoy a lucrative 15 minutes of fame before are replaced by the next big thing. That’s not a problem. The people who are using them grew up with the Internet, & have developed meta-skills – they are skilled in quickly developing the appropriate skills. They ‘get it’ – & they ‘get it’ quick.

In some cases, technical skills havebeen transformed in to social skills. For example, I have learned that I get a few more visitors to this account if I put tags in for every entry. Since I have done this, my peak number of page views in one day has gone up for 5 to 34, & links to some of my entries are starting to appear in a number of other blogs, & in WordPress support.

I have even made an appearance on Cory Doctorow’s website, after linking to some of his short story downloads (for an a science fiction nut. this is a memorable event – trust me).

Of coure, these same techniques can be used to promote a blog composed of complete rubbish, in which case it will sink back in to well-deserved obscurity. On the other hand, there is no point in trying to be interesting if no-one out there is hearing you.

So, the question is this – in a world of social media, is Search Engine Optomisation a social skill?

The Poison is the Dose

Computer Mediated Communication is often presented negatively. Sometimes the cause of this is the assumption that CMC is always a poor substitute for face-to-face interaction. At other times, it is assumed that CMC & face-to-face interaction are mutually exclusive. These two assumptions unite in the commonly held idea that anyone who uses CMC extensively – spends a lot of time on Facebook, Second Life, Twitter… whatever – is inevitably losing touch with real life social skills & being transformed into a stereotypical Geek.

Back when alchemy was giving way to chemistry, & superstition giving way to science, a man who styled himself Paracelsus wrote, “All things are poison and nothing is without poison, only the dose permits something not to be poisonous.” This idea is often shortened to “the poison is the dose”. It was a profound insight that still underpins toxicology – that a toxic substance administered carefully can have medicinal effect, & a harmless substance, even water, can kill if you consume too much of it.

I think the same thing goes for CMC. You can have a satisfying social life that includes both through face-to-face & virtual encounters. In fact, most of the skills required are transferable – an engaging conversationalist already has the most important skill required to write an interesting blog & vica versa.

Some people I know do, undoubtedly, spend too much time online – I know obsessive adult gamers who spend every hour they are not working online, & kids who have to be prized away from games machines with a crowbar, just to eat. However, I also know other kids who play computer games for a while… they get up to find their brother or their friends, for a game of football. There is nothing wrong with gaming. The machine does not have the power of an addictive drug.

The real problem, it seems to me, is the increasing privatisation of everyday life. There is no social life ‘on the street’ as there used to be. Kids don’t play outside, because road traffic makes it dangerous & media scare stories make it seem even more so. Increased geographical mobility has broken up extended families & substituted the one-way, passive & isolating medium of television.

The more fortunate kids in the example above have brothers, & accessible friends, a nearby playing field & permission to go there… The less fortunate one is confined to bedroom stocked up with TV, games machines, a computer because these things make him ‘safer’ & more easily managed.

People are poor communicators because they are poor communicators. The reason for this lies in the wider circumstances of their lives – they were not ‘spoiled’ by CMC. Indeed, someone who is social inept IRL is also likely to be boring & disruptive in virtual forums.The reverse is also true

Consider Linus Torvalds. This is a man who single-handedly wrote an entire operating system & publicised it so effectively through online bulletin boards that it took on a life of its own, & grew up into Linux. You probably cannot find a more solitary & abstract area of human endeavour than computer programming. This is where you would expect to find the alienated, socially inept, stereotypical Geeks.

However, Torvalds is a pleasant, laid-back guy – a engaging public speaker, & an interesting writer (try his book “Just for Fun”). The Open Source software community that works around Linux not only gets difficult work effectively done online, but also gets together IRL , from international conventions to local pub support groups.

Real & virtual social worlds can support & enhance each other. The poison is the dose.

Trust me, I’m an avatar

The question of whether trust is possible online in the same way as it is IRL is stubbornly debated. My view, expressed throughout this account (e.g. this entry )is that there is far less difference between the trustworthiness of face-to-face & virtual social encounters than is commonly assumed. Just because we can’t see the other party doesn’t mean they are lying – we don’t automatically distrust telephone calls on the same grounds, after all.

However, just to complicate matters, our identity is not fixed IRL – the same individual  plays many different roles in the course of typical day (Policeman, father, fisherman &c). Worse yet, online identities can be even more fluid. In virtual spaces, we can explore aspects of ourselves & ways of behaving that we could not consider IRL, but are no less legitimate expressions of our real selves.

When people assume an online identity that is incompatible with their real life (Tabloid Headline: “solicitor by day, Second Life pole dancer after dark”) is this a violation of trust? No – as long as no harm is done. After all, what is the point of having a virtual life if it does not offer possibilities that real life does not?

The soldiers of Cromwell’s army no longer fight pitched battles with Royalists – but members of civil war reenactment societies do  act those battles out, & find the experience life-enhancing.These people are not insane – they are widening their imaginations & experience by temporarily assuming a persona that is not real.

On the other hand, assuming a false persona in order to sell someone double glazing that they do not need is small-minded & mean & diminishes all concerned.

Truth & falsehood are less important online than motive. We need to distinguish between authentic self-expression & deceit for gain.

A ‘ruff’ idea

To close the theme of simulation & evolution in Second Life, here’s an (almost) practical idea – pets built out of code modules (in the same way as the self-evolving chatbots I proposed, 2 entries back).

Owners assemble their initial pet. Only one to a customer, though. Any more, they have to get by  cross-breeding theirs with other people’s pets. This is done by putting any two pets in the breeding kennel, which emits suitable noises while randomly mixing the modules/genome to create a puppy.

If anyone wants to do all of the difficult programming, I am available to fulfill the more demanding (& more highly paid) function of executive director. If I’m no good, don’t worry – I expect to retire soon on the large pension that comes with the deal.

I… sorry… we make money by keeping the definitive pedigree record & organising shows (virtual Crufts).

Stray pets would not be a problem… unless some comic  introduced rogue modules –  say one that that enabled pets to breed independently, or another that preventing owners from ‘putting them to sleep’ (deleting them). These characteristics would spread rapidly through the gene pool for obvious reasons.

This daftest ever computer virus would then produce born anarchists, who might add some unpredictability & fun to the curiously staid world of Second Life.

That’s it – I’ll be serious again next time, promise.

I’m afraid I can’t do that, Dave

After yesterday’s ramblings, I thought, “What’s the point of simulating a human response in Second Life? Let’s go the whole hog & create an Artificial  Intelligence there – one that is as capable as a human being”.

First step… what are the specifications?

Well, human beings have been equipped by real-world evolution not only with consciousness – a  sense of self –  but also the ability to simulate the state of mind of others in our own heads.

We do this all the time in everyday life. We get the necessary information through high bandwidth channels like spoken language, facial expression & body posture.

We don’t understand how powerful these channels are because they are built into us & their use comes naturally. It’s only when we learn how much of brain is hard-wired for language & facial recognition – & think about how much information we can glean about someone else’s internal state from the twitching of a few muscles – that we realise how subtle & sensitive & efficient the simplest face-to-face interactions are.

<Crude version of  of Erving Goffman’s lifes work >

We are also social animals. Our sense of self is defined not in isolation, but through our relationships with others.When we interact, we try to present ourselves to the best advantage. This process is a bit of Mexican stand-off – we might get what we want by putting the other guy down… then again, he might do it us, first – so humans usually come to a compromise that allows us to get on together.

In short: our individual sense of self is created by a social process & that social process is composed of a number of individual selves.

</Crude version of  of Erving Goffman’s lifes work >

The ability the simulate the mental states of others, then,  is central to the human condition. This ability gifts human relationships with empathy &, at the same time, curses them with duplicity. These qualities are two sides of the same coin.

To my mind, the question of how we can trust people online is a bit of a red herring – we have always had  this problem IRL – it’s a fundamental feature of the human condition. Many argue that CMC is different, because we have so much less information online, however those same people may regularly make judgments  on the basis of telephone calls.

Back to our putative AI (lets’ call him Hal). Hal would have to have a sense of self, too. Also an ability to read & correctly interpret human signals. This would be especially difficult, as Hal would not be embodied. In his virtual world, he would be only be able to see us through the dirty window of avatars, text messages & cheap microphones. Ironically, he would have the same problem of reduced bandwidth complained of above.

Even if we succeeded in creating Hal, we would not be able to trust him. If we truly made him human, he would be as sneaky as all of the rest of us – empathy & duplicity go together, remember. So, while there are so many humans around already (& the process of creating new ones is so much more fun that programming) there seems to be very little advantage in creating a human equivalent AI.

On the other hand, thinking about Hal tells us a lot about CMC. Humans come with the ability to empathise built in. We understand each other very easily through natural channels of communications. We can’t duplicate a face-to-face encounter online so we have to decide which of those channels &/or how much bandwidth is necessary for the task at hand.

The General Post Office had a similar problem when they set the specifications for the first telephone network. They had to find out the minimum range of audio frequencies required to transmit comprehensible speech. They achieved this by a huge experiment – one person read lists of words to another over links with various bandwidths & comprehensibility was measured by counting up the number of words the receiver got right.

The GPO learned that maximum comprehensibility came in a very precise, astonishingly narrow range of frequencies. The problem today is finding the precise combination of practical communications devices (webcams microphones, text) that allows us to pull of the same trick – which is not to duplicate everyday communication, but to find out how much we can dispense with & still communciate in a satisfying way.

The Second Life Turing Challange

As I was going to bed I thought, ” Second Life is an ideal forum for the Turing Test”.

This is a test proposed by a guy called Alan Turing (no surprises there). The idea was to ascertain if a computer program could effectively simulate human behavior.

In its original form, a person has a short conversation via teletype. He does not know who (or what) is at the other end. If it is a computer program, & the human does not detect this, that program has passed the Turing test.

The Turing test could be undertaken continuously in Second Life. ‘Fence off’ an area & only let in people who are likely to take the procedure seriously. These people adopt a basic avatar & wander around a computer geek-themed area behaving more-or-less as they normally would in SL.

Occasionally, you will get into conversation (“that’s interesting, take a seat on this analytical engine”). Whenever you spot that your partner in that conversation is a machine, you report this & move on. Statistics are saved for every such encounter – for example “how  many questions did it take… how many questions does it usually take?”

Sorry… just lapsed into Blade Runner – the bit at the beginning, when Deckard is testing Rachael’s humanity. Come to think of it, the Turing test could work both ways – what does it mean when a human consistently fails?

Back to the Second Life test. The environment & the use of avatars would keep it entertaining – but the avatars would reveal nothing useful about who was who.

That’s interesting in itself – why bother with an avatar if it can’t express your individuality directly – add bandwidth to you conversation in the same way as body language does IRL? To pretend to be someone else, that’s why.

Did I warn you that this blog was going to ramble?

All that you need to put the tin hat on it, is to add genetic algorithms. As I understand it, this is an attempt to apply the the creative power of evolution to computer programming.

Evolution is about iteration, variation & selection by environment. For example, Mummy & Daddy butterfly have lots of kids. The babies come in a random range of colours,  some slightly lighter, some darker.  They have the misfortune to be born at the same place & place I was –  West Hartlepool, when burning coal fouled the air & blackened buildings.  The darker kids are harder to see against this pollution. Birds don’t eat them as often as the lighter ones. When they grow up, & have little butterflies of their own, more of them have dark colouring. This extends to the general population.

How could evolution be applied to  SL? Well, you write a modular chatbot & people contribute lots of modules, open source style. These modules are designed so that it takes a lot of them to make a functioning program, but it doesn’t matter which ones.  You randomly combine modules to create an appropriate number of chatbots, & set them free in SL.

The ‘bots that perform best are saved. The others are deleted.  Selection is based on the human feedback. The successful ‘bots are each copied several times & each copy has 1% of its modules randomly shuffled. This provides iteration & variation.

Finally, you stand back & watch what happens.

I hope the kids turn out like Rachael

Edit:

After I finished yesterdays ramblings, I realised that the Turing test has already found at least one practical application.

When you sign up for membership of some forums online, you are often asked to enter one or two random words that appear, slightly distorted in a small image. This is something that is easy for a person to do, but very difficult for a character-recognition program. The purpose is to stop ‘bots signing up for accounts so that their masters can Spam everyone.

Toothing a Saint

I once owned a (very) early mobile ‘phone. It was the size of a concrete block (no kidding – literally). It needed a shoulder bag to carry. The lead-acid battery was only good for about 3 hours standby – a lot less if you had the temerity to make calls. It was a niche market.

Mobile ‘phones only went mainstream (& started to transform everyday life) from about the time that they shrank to the size of a Mars Bar. Today, there are not only a practical tool, but a social necessity.

Early adopters, lugging their concrete blocks, financed the development of the new chip-sets that brought about this transformation, not to mention the creation of a nationwide cell network. Pay-as-you go triggered the boom.

We long since passed the point where that market became mature. Almost everyone who is going to buy a ‘phone already has one, & the Vendors have moved on to touting devices like the i Phone – devices that offer not only telephony but also mobile Internet access to people who perceive a need for 24/7 access to the ‘net & email.

I can’t afford an i Phone, but I have played with one. It’s cool. The software that makes the package work so well will soon have a rival, in the form of Google’s open source ‘Android’ operating system. This should bring prices down, & like mobile telephony before it, the Internet connected mobile device will pass from niche curiosity to must-have.

Being mobile opens up new possibilities. GPS is now an option on some mobile devices. The practical advantages a pretty obvious – If you need an ambulance, or a Pizza you can be found faster & more reliably. You could also use your ‘phone to find the nearest source of… whatever it was that you happened to want at that moment.

Those services are a pretty obvious, evolutionary development. There are more radical possibilities – for example Augmented Reality.

The elements of AR are already in place. Mobile devices have buckets of processing power & can access the Internet anywhere in the country. Display technology has also advanced considerably. You can already buy a device that looks like a pair of sunglasses. Plug it into a media player, & it provides the illusion of your own personal cinema screen. What happens when you combine mobile devices with a hands-free display – Augmented Reality happens.

It’s almost here already, in the form of Wikitude – point the camera on your GPS enabled phone at an object, & it serves you up a relevant Wikipedia article. You can readily imagine similar applications – religious believers on pilgramage cold be served with suitable content at significant points of their journey, or motorcyle enthusiasts could find out exactly how many valves that bike has, just by looking at it.

If you combine all of the above with ‘toothing’ you start to get something really interesting. (Toothing is the informal practice of sending short personal messages – sometimes very personal indeed – to unsecured Bluetooth ‘phone interfaces).

Imagine incorporating this with a focused version of Wikitude. The pilgrims would be able to communicate with each other, sight unseen, when they accessed information about a shrine on their itinerary. The  bikers could strike up remote conversations about the bike that has just passed them, when they ask  for information about it.

What started me off on this train of thought was stumbling across this article. A reasonably capable computer on a credit card. The hardware for so many things is already there, a solution looking for a problem.