ANONYMOUS SOCIAL STAMPS

Authenticating Anonymous Statements to Friends with Privacy

Sandeep S. Kumar

Information and System Security Group, Philips Research Laboratories, Eindhoven, The Netherlands

Keywords:

Privacy, Trust, Social network.

Abstract:

Numerous online services and applications for smart devices exist today on the internet which claim almost

similar functionalities. This inevitably leads to the problem of being able to choose the right one which is

mostly done today by trial-and-error. However this gets tricky if it involves sharing privacy sensitive infor-

mation with the service e.g. in the case of health and well-ness services. A customer’s trust in a new online

service is known to increase based on testimonials (or observable behavior) of people within a social distance

(friends, friends-of-friends etc). Linking testimonials to users requires releasing one’s social network infor-

mation which by itself is privacy invasive. The paper presents the concept of an anonymous social stamp

which can be assigned to anonymous statements and help prove that the statement was made by a particular

member of an external social networking site. Trust can then be derived based on the social distance between

the persons concerned: the person who made the statement and the person verifying the statement. A possible

implementation of the concept is shown with the existing infrastructure of RSA keys already in use by social

networking sites. The concept is applicable for the new services that can create conﬁdence in new users by

revealing anonymous data sharing conﬁgurations of other users with the service.

1 INTRODUCTION

An incredible amount of choice exists today in web

services and applications that can be used online or

downloaded onto various smart devices. With this

incredible choice comes the problem of determining

which of them are useful and actually deliver what

they claim. These services and applications can range

from simple productivity applications like collabora-

tive text-processing to more advanced applications for

health and well-ness like tracking vital bodily sig-

nals. Most of these applications require sharing cer-

tain amount of privacy sensitive information. This

makes it evenmore important to make an educated de-

cision on which applications to install and how much

information to share with such services.

It is well known that people tend to use testimo-

nials and advices provided on various aggregation

sites like epinions (Epinions, 2010) while choosing

services. The main idea behind such reputation sys-

tems (Resnick et al., 2000) is that the majority of the

crowd can help ﬁlter out scrupulous service providers.

However the concern is that such systems are easy to

manipulate since ratings are provided by people with

no prior trust relationship. Therefore most of these

trust and reputation systems (Josang et al., 2007) are

faced with the problem of linking pseudonyms to peo-

ple one can trust. Another concern is that such meth-

ods do not help identify services that would best ﬁt

to one’s individual needs but rather on average to a

wider group.

Health and well-ness services is one of the ar-

eas that urgently require trusted statements and help

in making educated decision of using different ap-

plications. Health and well-ness has become a buz-

zword that is increasingly being used by many fraud-

ulent services to attract new customers. Such services

mostly don’t work or are sometimes even not safe for

some consumers. The increasing numbers of such

incidents have even lead the Federal Trade Commis-

sion (FTC, 2010a) to set up a dedicated information

site for consumers to detect misleading and decep-

tive health claims at Who Cares (FTC, 2010b). The

European Healthcare Fraud and Corruption Network

(EHFCN) has recognized that one of the reasons that

fraud happens is when providers could have conﬂicts

of interest that affect their judgement (EC, 2010a).

Customers that have either experienced or heard/read

about these much publicized frauds are very reluctant

to use any new heath and well-ness services. There-

183

S. Kumar S. (2010).

ANONYMOUS SOCIAL STAMPS - Authenticating Anonymous Statements to Friends with Privacy.

In Proceedings of the International Conference on Security and Cryptography, pages 183-188

DOI: 10.5220/0002993401830188

 SciTePress

fore, creating trust in the new heath and well-ness ser-

vices would be a very important requirement for new

service providers. Additionally, customers need to be

conﬁdent to share vast amounts of sensor data and pri-

vate medical information that would be needed to pro-

vide the service. Legislation like the Unfair Commer-

cial Practices Directive (Directive 2005/29/EC) (EC,

2010b) can only deter certain amount of misleading

claims from service providers but cannot create con-

ﬁdence in the services as being beneﬁcial.

1.1 Our Contribution

Knowledge about the social distance of a new cus-

tomer with existing consumers of a service is impossi-

ble to determine without access to the social network

of the new and existing customers. However, this data

is readily available at the numerous social networking

sites. These social networks have evolved into an im-

portant source of communication and trust building

on the internet. These trust relationships can either

be the individuals trust relationships in the physical

world or new ones based on virtual interactions.

For a health and well-ness service provider, cre-

ating and administrating yet another social network

would not be preferred option but to have a mecha-

nism that allows tapping into the existing information

at the social networks to give the assurance of the so-

cial distance between two parties. However, revealing

a person’s static identity at a social networking site or

openly revealing one’s social network is privacy inva-

sive. Therefore external sites that would like to create

trust in their services through recommendations from

friend’s network, need a mechanism to link into the

social networks without invading privacy. The mech-

anism should provide enough ﬂexibility such that in-

dividuals are not forced to reveal their social networks

openly if they do not desire. Additionally, the social

networking site should not be able to gather and learn

about the various external services that the individual

subscribes to based on this mechanism.

The main idea of the paper is an anonymous social

stamp (created using existing RSA key certiﬁcates of

the social networking sites) that individuals can attach

to a statement (reviews, advices, conﬁgurations, etc.)

on service providers web-site or other review aggre-

gation sites for others to view. A new customer of the

service or application can view this statement and if

within a social distance can verify it.

The paper is organized as follows, in Section 2 we

clearly deﬁne the setting and the various parties in-

volved. In Section 3 the security and privacy require-

ments for our solution is described. Section 4 presents

the construction of our anonymous social stamp solu-

tion. The various interactions for initialization, gen-

eration and veriﬁcation of the stamp is also presented.

In Section 5, we present other useful applications of

the anonymous stamp and end with conclusions in

Section 6.

2 SCENARIO

As mentioned previously, testimonials and advices to

use new services can have additional value if it can be

proved that they are written by someone within a so-

cial distance like friends-of-friends. This would pro-

vide some amount of trust in these statements com-

pared to statements made from total strangers. We

use a social networking site as a way to determine

this social distance. To make clear the setting, we

ﬁrst present the various parties in the system and their

roles. Then we discuss the security and privacy re-

quirements for the different parties involved.

2.1 System Model

We present the relevant parties with an example as

shown in Figure 1. Alice is the party providing an

anonymous testimonial (or rating) of a service. How-

ever she would like to attach a proof (stamp) that can

help others viewing the testimonial to derive some

amount of trust in that statement, if they are related

within a social distance. Bob is a new customer who

would like to use the service and make an educated

decision based on the testimonials (one of which is

from Alice). He retrieves the testimonial and uses

the anonymous social stamp to contact the social net-

working site to know if the anonymous testimonial

writer is within a certain social distance of him. Based

on this information he can decide to either trust the

testimonial or ignore it.

3 SECURITY AND PRIVACY

REQUIREMENTS

For the anonymous social stamp to work as shown in

the scenario it has to fulﬁll certain security and pri-

vacy requirements as mentioned below:

• Veriﬁability. The social stamp can prove to a ver-

iﬁer (Bob) with the help of the social networking

site that the statement indeed belongs to a speciﬁc

user of the social networking site. The real iden-

tity or pseudonym at the social networking site

is not revealed but only the social distance is re-

vealed by the social networking site based on the

SECRYPT 2010 - International Conference on Security and Cryptography

184

Figure 1: Implementation scenario for the anonymous social stamp.

veriﬁer’s (Bob) social distance to the statement

creator’s (Alice) at the networking site.

• Secure Binding. The stamp is publicly visible

to everyone and therefore should not be possible

to copy to a different pseudonym or statement by

anyone.

• Privacy of the Creator. The stamp by itself does

not reveal the identity of the creator (Alice) to

anyone except to the social networking site.

• Unlinkability of the Creator. The stamps alone

do not enable linking the creator (Alice) across

different sites.

• Privacy of the Veriﬁer. The veriﬁcation of the

stamp does not require the veriﬁer to show the so-

cial networking site the statement or the site from

which it was obtained. This is minimize the so-

cial networking site from knowing the different

services he is interested in.

4 THE ANONYMOUS SOCIAL

STAMP

Before we present the technical details of the Anony-

mous Social Stamp, we ﬁrst assume the following ter-

minology:

• The external service can be uniquely identiﬁed as

W. This could be the unique url or the SSL cer-

tiﬁcate of the service.

• The statement (review, advice, etc.) to which

the social stamp needs to be attached at W be

represented as T

, and uniquely identiﬁed by the

pair T

,W. For practical purposes, we assume

that the pair T

,W can be converted to a single

value h using a publicly known hash function, i.e,

= Hash(T

,W).

• The social networking site is represented as N.

• The user that creates the social stamp (Alice) has

the pseudonym I

at the social networking site

N. She also has a shared secret S

with social

networking site (e.g. a hashed password).

• The veriﬁer’s (Bob) identity at the social network-

ing site N be V

ANONYMOUS SOCIAL STAMPS - Authenticating Anonymous Statements to Friends with Privacy

185

• The social network N possesses the RSA public-

private key pair {P

}. Based on (Rivest et al.,

1978), the keys P

and S

exhibits the following

property:

)

mod n = g (1)

for any g and for public n as deﬁned in (Rivest

et al., 1978). The method to generate these RSA

keys and their properties are well known and will

not be described further here. However note that

all operations shown further are done mod n

even when not explicitly mentioned.

We now describe how the anonymoussocial stamp

can be created for a statement and veriﬁed. Figure 2

shows the various phases from initialization, genera-

tion and veriﬁcation of the stamp. We describe each

step in detail and show the correctness of the protocol.

4.1 Initialization Phase

The ﬁrst step is the initialization phase where the

user Alice at the social networking sites is provided

with keys to create any number of anonymous social

stamps. The social networking site generates a large

random g. This g is ﬁxed for each social network and

known publicly. Alice who possess the identity I

the social networking site, is given the two keys K

and K

derived as follows:

• K

= g

, which is same for all users and need

not be kept secret since it is created using a public

key.

• K

= g

)

, which needs to be kept secret by

Alice. However Alice can verify that it indeed is

key based on her shared secret decrypting it using

the public key P

The social networking site stores the pair {K

}

in a lookup table. This is useful to perform a reverse

look-up of the user identity from K

4.2 Stamp Generation Phase

Now, if the user Alice wishes to attach a social stamp

to a statement uniquely identiﬁed as {T

,W} on an

external site, then the following steps are perfomed:

1. Create the hash h

= Hash(T

,W).

2. Generate a large random number r

, keep it secret.

3. Create the Binding stamp: B

= K

)

4. Create the Helper stamp: H

= (K

)

5. Assign the tuple {N,B

} as the anonymous so-

cial stamp to the statement T

,W.

Note that the stamp is bound to the statement only

in such a way as to prevent copying the stamp on to a

different statement. It however does not prevent any-

one else to copy the statement and create their own

stamp for it. This should however not affect the secu-

rity requirements since our intention is to verify if a

statement can be linked to Alice, irrelevant of the fact

if she had copied that statement from someone else.

4.3 Stamp Veriﬁcation Phase

When a new customer Bob would like to verify the

stamp {N,B

} attached to the statement {T

,W}

then he should be an existing member of the social

networking site N. Assuming he too is a member at

N with the identity V

, he can perform the following

steps to verify the stamp and get the social distance:

1. Create the hash h

= Hash(T

,W).

2. Transform the helper stamp so that the veriﬁer

Bob can be sure that the stamp is indeed bound

to h

by performing the following operation:

= H

= g

.r.h)

3. Now contact the social networking site N with the

tuple {V

} to verify the binding and get the

social distance.

On receiving the tuple {V

} the social net-

working site performs the following operation

1. Use the RSA secret key pair of S

to generate a

de-binder from the new helper stamp as follows:

= (K

)

= (g

)

= (g

)

= g

)

The result is due to the property of the RSA keys

as shown in Equation 1.

2. Use the de-binder

to remove the statement re-

lated hashed values from the binding stamp by

performing:

= (K

)

)/g

)

= K

This operation also veriﬁes that indeed the bind-

ing was for the statement h

it was intended for,

otherwise the generated key will not be right.

3. The social network can now perform a reverse

user identity lookup for K

in its lookup table

to map to the identity I

of Alice as the creator of

the stamp. The social network neither knowswhat

the statement was about or where it was posted.

SECRYPT 2010 - International Conference on Security and Cryptography

186

1. Assumptions:

• Communication channel between all parties are authenticated (for e.g. using pseudonym and password)

• All operations performed mod n as in Equation 1

2. Initialization Phase:

Alice: pseudonym I

Social Network N: RSA key pair {P

}

Shared secret: S

Generate a random g (public)

}

 K

= g

)

Secret keys {K

} Store pair {K

}

3. Anonymous Social Stamp Generation Phase:

Alice: creates statement T

Service Provider W

= Hash(T

,W)

Generate random r

(secret)

Binding stamp:B

= K

)

Helper stamp:H

= K

,{N,B

}

- Publishes statement: T

Social Stamp: {N,B

}

4. Anonymous Social Stamp Veriﬁcation Phase:

Bob Service Provider W

Retrieve: T

,{N,B

}

,{N,B

}

 Published: T

,{N,B

}

= Hash(T

,W)

= H

Bob:pseudonym V

Social Network N

Authenticate to N

Verify bind and social distance

,{B

}

= K

Reverse lookup K

→ I

Distance d between V

and I

Bind: Yes/No and distance d



Figure 2: Anonymous social stamp interactions.

4. The social network can now inform the veriﬁer

Bob (after he has properly authenticated to the so-

cial network), the social distance between I

and

This social distance allows Bob to assign a certain

amount of trust to the statement rather than assuming

it to be a complete stranger. Alice does not have to

worry about revealing her interactions at other sites

to the social networking site or to be linked across

the different sites based on her stamp. Thus all the

security and privacy requirements laid down before

can be met using anonymous social stamps and still

provide the functionality that is required.

More advanced features can be built around this

main idea in which the social network allows users to

classify users to be more trustworthy only if they are

connected through a branch of a particular friend, etc.

ANONYMOUS SOCIAL STAMPS - Authenticating Anonymous Statements to Friends with Privacy

187

5 ADDITIONAL APPLICATIONS

Here we present an extension of the above social

stamp concept where the main aim had been to au-

thenticate that a statement had originated from a per-

son in the social network. However, the technique

can also be used if the user wants to hide information

which can only be read by people within a social dis-

tance. Thus additional information can be revealed if

the person viewing the statement is within a certain

social distance.

To enable the hiding of information, the statement

needs to be encrypted with the key Key = K

)

Then the same steps for generation of the stamp is

followed as before for the pair {T

,W} .

During veriﬁcation of the stamp, the social net-

working site performs all the steps as before to iden-

tify the user as I

. Once the user is identiﬁed, the

social network knows the shared secret S

and there-

fore can also re-create the key by performing the fol-

lowing operation:

Key =

)

= (g

)

= (g

)

= K

)

The user I

can conﬁgure the minimal social dis-

tance that needs to be satisﬁed for the social network-

ing site to release this Key. If V

is within this social

distance then he receives the Key from social network

and decrypt the message T

6 CONCLUSIONS

We presented the concept of an anonymous social

stamp which can be assigned to anonymous state-

ments and help prove that the statement was made by

a particular member of an external social networking

site. Trust can then be derived based on the social

distance between the persons concerned: the person

who made the statement and the person verifying the

statement. A possible implementation of the concept

is shown with the existing infrastructure of RSA keys

already in use by social networking sites.

This should enable a serious health and well-ness

service provider to show testimonials, advices and

conﬁgurations which can be veriﬁed by new cus-

tomers to belong to people that they would consider

trustworthy based on their social distance within a

particular social network of their choice. Customers

can use any social network of their choice and trust.

Service providers would be able to use the trust that

people have created by the social networks without

having to create a social network of their own.

Additionally we extend the concept to hide infor-

mation that can be decrypted only by people within a

social distance.

REFERENCES

EC (2010a). European healthcare fraud & corrup-

tion network (ehfcn). http://www.ehfcn.org/fraud-

corruption/fraud-and-corruption-in-healthcare/.

EC (2010b). The unfair commercial practices directive di-

rective 2005/29/ec. http://ec.europa.eu/consumers/

rights.

Epinions (2010). http://www.epinions.com.

FTC (2010a). Federal trade commission: Health claims.

http://www.ftc.gov/bcp/menus/consumer/health.shtm.

FTC (2010b). Who cares? sources of infor-

mation about healthcare products and services.

http://www.ftc.gov/bcp/edu/microsites/whocares/

index.shtml.

Josang, A., Ismail, R., and Boyd, C. (2007). A survey of

trust and reputation systems for online service provi-

sion. Decis. Support Syst., 43(2):618–644.

Resnick, P., Kuwabara, K., Zeckhauser, R., and Friedman,

E. (2000). Reputation systems. Commun. ACM,

43(12):45–48.

Rivest, R. L., Shamir, A., and Adelman, L. (1978). A

method for obtaining digital signatures and public-

key cryptosystems. Communications of the ACM,

21(2):120–126.

SECRYPT 2010 - International Conference on Security and Cryptography

188