Reputation Economies in Cyberspace - Yale Law School 12/8/07

Reputation - "its more about relationships than transactions" "it’s about authenticity"

Shia from IBM: Mingle outside, that’s where a lot of the more interesting things happen.

We’ll put up a wiki and photos of this.

ISP.blog.yale.edu

What’s at stake?

Not just economically, but from a normative standpoint?

What do we make from the shift from closed systems of identity verification to more community based systems of regulation.

Questions of informational integrity, and of who owns identity.

Should this be handled privately, regulated, should there be an NGO to deal with it?

Michel Bauwens. Operates from his home, a global cybercollective: the Peer to Peer Foundation, to study p2p dynamics. The technical organization of society is shifting to more decentralized, bottom up structure. Peer governance, self management, the ability of institutions to protect information.

As Yochai Benkler says - nonreciprocal engagement... every physical production has a design phase & there’s no reason this can’t be done in the same distributed, decentralized manner.

There is already a thriving community of communities - designing furniture and other physical goods; they strive to be not just better than competition, but to make the best thing in an absolute sense... this eliminates planned obsolescence you find in industry.

3 Business Models emerging. Each w/ it’s own social contracts..

1. Sharing Economics - individuals wanting to share creative expression. Weak links amongst each other. Less 1 to 1 engagement. Different motivations of the sharers vs. proprietary motivations. No crowding out. People seeking money are in the minority. Once you start integrating revenue into that, it changes the community. It’s wrong to think you can monetize a sharing economy. Esteem, reputation, intention are not monetizable in this kind of economy. The logic is different.

2. Commerce Oriented Production. People working together to make things. One variant is commons oriented peer production. They’re not using proprietary platforms, but ‘for benefit’ platforms. These for benefit institutions are not interested in the monetary aspect. Still depends on voluntary work. So what do they do with money?

3. Crowd sourcing mode. Partial integration of P2P dynamics in the corporate value chain.

Tradeable goods

Measurable goods

Some digital goods have relative value only in their specific contexts.

Find wealth acknowledgment systems that acknowledge contributions to the community, even in the non-business context.

MB argues against reductionism of non-market systems to market systems.

Rishab Ghosh - United Nations University - MERIT:

everything in traditional economics: pricing is only a measure - in the real world, other information is necessary to make decisions - info not encapsulated in price alone.

The big difference between informal economies & formal economies is the pricing.

How do you select which open source project to work for? Price isn’t a part of the equation. Price doesn’t encapsulate everything you want to know. (VALUE!!!-mp)

It’s natural to think of reputation as a convertible currency. It’s convertible differently for different people. But it’s not a straight line exchange rate - ‘winner take all’ exchange rate spreads apply.

That’s the situation we have today. The structure for converting isn’t evenly distributed.

Online Reputation today:

e.g. John runs restaurant, some like it, some don’t. Everyone rates it. It’s good food, but gets bad rating, because most don’t like that kind of food. - this doesn’t happen so much in the real world.

E.g. Book recommendations - one from a trusted individual is very different from the systems online.

E.g., systems are easy to abuse e.g. people who don’t like French food, who will therefore pan the french restaurant. Sock puppets (planting stories to pump yourself up). In the real world, there’s no objectivity of opinion. Online reputation systems make the mistake of treating reputation as objective, which is wrong. Online systems need to be more like real world.

Auren Hoffman, of Rapleaf.

Consumer Control in a Data Driven World.

1. Data on you is growing - DOB, SON#, CIA/NSA/FBI, Credit history, Web footprint, genome.

And the RATE of growth of data being collected on you is also growing.

And the number of actors collecting data is growing: sellers, financial institutions, government, advertisers, .... if everybody has a bomb, there’s more of a chance that something bad will happen.

We live in an opt-out world. Most collect data on you, like it or not.

In the opt-in world, user decides. Challenge is that there are few users, hard to get everyone on board. Look at grand failures like Microsoft Passport, that required high adoption rate for critical mass.

In the opt-out world, you get critical mass, but at the expense of privacy. Other benefits of opt out world. Readily accessible credit. Micro targeting. Cookie enabled websites.

Look for consumer action and/or government action to strike a new bargain, and soon.

Consumer Data Rights:

1. To find, own and share your data.

Accessible, Portable, and Changeable.

No one should know more about you than you know about yourself. You should be able to go to doubleclick, NSA, etc. and see what they have on you.

2. You should be able to change incorrect information. Hide private information.

3. Portability - your profile on one social network should enable you to fill in elsewhere.

Privacy is not easy.

Giving people the opportunity to quickly opt-out.

Many gray areas: MIT case study. A mother in search her missing son, had to hire a private investigator after the school denied her access to son’s room, Facebook account, etc, on privacy grounds. The son had been killed.

auren@rapleaf.com 657 Mission Street, suite 600, San Francisco, CA 94105

- rapleaf lets you plug in your e-mail and get some kind of reputational score.

Problem is systems of reputation online don’t capture reputation so well as real world.

Reputation is very context dependent. Love the fico score - it measures one thing alone (credit risk).

Hassan Masum - McLaughlin Center for Global Health

"Making your name online"

when is there enough time? Never

Is x really true? Why do you believe me?

Examples

epinions.com

slashdot.org, Kuro5hin.org

Ryze

Advogato.org affero.org

TrustRank, HITS

Are online names most important...

To enable distributed collaboration?

To complement mass media POV..

To search thru masses of info

to block spam and other badness

to keep up in the Red Queen’s Race?

How do we aggregate trust? Trust networks.

A node is a user. An Edge is a trust statement. Trust Metric (TM) predict values of trust

e.g. google: nodes are web pages, edges are links. Google computes ‘importance’ of every page similarly...

would you rather be highly rated by a small group of peers or the world at large?

In which areas are we all experts?

What does distribution of domains look like

Our democracies are only as strong as our awareness, insight and time to participate. ***

How can democracy scale? "Dynamically Distributed Democracy"

What would it take to amplify the most thoughtful and insightful voices?

What about for unpopular points of view?

Which reputations are Stable?

Greater agreement: Nobel winner... BBC site...

Less agreement: pop singer... divisive leader...

What Models to study - what "canonical models" could offer insight to a broad community?

Propagation of Trust

* Six Degrees by Duncan watts

* Critical Mass by Philip Ball

Time and Number:

Real-time: talk, chat, TopCoder

Long-term: books, ‘cumulative blogs’

Beth Noveck

was asked by someone who was being told by Sony they had to give up their identity and points in Everquest.

Legal discussion has been mostly about managing risk, authentication. Many of us are relying on wiki production - not just ‘love’based, but to create value with virtual companies.

Legal doctrines of transparency, portability, etc. Look to the role that groups and communities can play. There’s a technological and economic opportunity here.

- look at Amazon. Useful, not a lot of teeth.

- LinkedIn model -

- resume model

Getting social groups to pledge to back their members’promises

Blog posting about "Micro elites" - small group production - in many contexts, all we need is the assurance of that small group, perhaps backed by real financial teeth.

Implications for /application to democracy... Back participation with reputation.

Moderator: Is it possible to opt out? Google yourself... you don’t control. People write nasty things.

Reputation systems in some instances have to take into account zero sum consequences. Your reputational score is, in some cases, a relative thing.

RapLeaf is not - it’s just a score indicating probability of loan repayment.

Other systems are different.

Michel Bauwens - wikipedia last year chose a role of artificial scarcity, enforcing rules where they had not been before.

Q: what about forgiveness, after someone repairs their credit/trust score

**Hassan: if you don’t forgive, you miss information.