Ethics and the Use of Sensor Network Monitoring in an Art Setting Maarten Hoek University of Twente P.O. Box 217, 7500AE Enschede The Netherlands

[email protected]

ABSTRACT This paper deals with the use of sensors at a museum or art exhibition. Privacy issues are considered by surveying the public about their opinion and by using previous conclusions from related work. Using previous works in this field and the state of the art, three possible uses for sensor networks in museums are proposed: analyzing visitor movements, an advanced security system and an interactive installation. The privacy implications for these examples are analyzed, using general properties of such a system: Resolution, personal identifiability, utility, relevance, pervasiveness, property and retention time. Visitors considered being personally identifiable the biggest breach of their privacy. There’s a lot of potential for interative installations in museums. Museums, artists and visitors all consider these interesting. These installations can be designed to not store any captured images, so privacy isn’t breached.

Keywords Modern art, sensor network monitoring, privacy, survey, museum

1.

INTRODUCTION

Sensor Networks are playing an increasingly important role in everyday life. For example, your smartphone might contain a GPS-sensor that can track your location every second of the day. Sensor networks in public and semipublic spaces allow for detailed analysis of everything that passes by. There are even new advances in Telemedicine that allow medical staff to monitor a patient’s health from a distant location. These new technologies can help society in new ways, but can also invade the privacy of the people that are being monitored [6]. Museums are also using sensor networks in exhibitions for different purposes [2] [9]. These new applications of sensor networks might be a threat to the visitors’ privacy, like many new technologies. How can we find out what is ethical in that setting? Privacy seems an important issue, but in practice most people have no problem with giving out their personal inPermission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. 15th Twente Student Conference on IT June 20st , 2011, Enschede, The Netherlands. Copyright 2011, University of Twente, Faculty of Electrical Engineering, Mathematics and Computer Science.

formation to sites like Facebook. Research [3] shows that most people simply do not care under these circumstances. It is therefore logical to research the opinion of the general public, if we want to find a useful conclusion.

1.1

Research questions

This research will try to find uses for sensors in museums that are both practical and ethical in terms of visitor privacy. The following questions will be answered: • What are the practical uses of monitoring visitors at an exhibition? • What do visitors consider ethical when they are being monitored? The views of visitors might change depending on the circumstances. The opinion of the visitors may be heavily influenced by the type of exhibition. Different exhibitions attract different types of visitors. Their opinion might also be influenced by the presence of the sensor network itself, especially if the monitoring is being done in a confronting and visible matter. Therefore these two questions have to be answered: • Can the type of exhibition influence a person’s views on privacy? • Do the visitors’ views on privacy change if they know they have been monitored?

2.

STATE OF THE ART

There has been a great deal of progress in the field of using information and sensor technology in museums the past years. This chapter will reference a few projects in this field to show what can already be done. A lot of museum-based technology revolves around enhancing the experience of visitors. These serve as a hightech implementation of the audio tour. Sparacino [12] describes such a system that augments reality to show relevant information about the art pieces. While these technologies are certainly interesting, they are opt-in by definition, since the user has to agree to wear such a device. They are therefore not interesting for a privacy research. This paper will thus focus on systems that are present in the surroundings. It is also possible to use sensor networks for measuring the environment of exhibition rooms [2]. There should not be any privacy issues with such a setup, since no information about people is collected.

2.1

Proposed uses

Based on the literature research, this section proposes three possible uses of sensor networks in museums. These will not be implemented, but will instead serve as examples.

Utility and Relevance Any system being deployed must naturally have some usefulness for it’s environment. Visitors might accept a breach of privacy of they get something in return. The purpose of the system must also have some relevance to the setting, or it will feel out of place.

Tracking Visitor Movements A system that locates visitors and tracks their movements around the exhibition or museum space. The data will then be analyzed to find out which piece was the most relevant, and how visitors move around the exhibition space. This information can be used to enhance the museum experience by giving information about which placement of pieces gives the best experience. An implementation of such a system was made by Cornell University at the Herbert Johnson Art Museum [9]. This sensor network can collect and process practical data about the visitors in the museum space.

Pervasiveness and Visibility A museum is designed for visitors to have an experience with art, so any system must not distract from the art in the museum. Secondly, visitors might feel uncomfortable seeing cameras all the time. Pervasiveness describes here how many sensors are placed. A system with just a few cameras is not very pervasive, a system with a lot of cameras and other sensors is very pervasive. More pervasiveness might increase the perceived privacy breach, but can give the system more information, since there are more sensors.

Security Sensor networks can be used to build a smart security system. Such a system might automatically detect suspicious behaviour based on the sensor input. This can protect the museum from vandalism and theft. Such a system was already made for the military [11], it might be possible to use a similar system in a museum.

Interactive Art Sensors can be used to create installations on the edge of art and technology. These installations can use sensor data to present data in an attractive way or create modern art that changes directly based on the input of the sensors. Artists are already using sensors to create experimental art. One Million Heartbeats [4] is an example of an artwork that uses sensor technology to really capture the participants. The experience of this piece is directly influenced by the sensor data and the actions of the viewers. Boehner and Sengers [1] describe an installation called Birdscape, a project that stands between art and tool. Museum Twentse Welle in Enschede currently uses sensor technology in an interactive wall that represents the history of Enschede [5]. The images on the screen change based on the location of the viewer.

3.

THEORY

As with any system that collects data about people in a semi-public space, sensor networks in museums can have privacy issues. There are several properties of sensor networks that influence privacy issues [8] [6]. This section will name these properties and determine if they raise issues for the proposed uses of sensor networks in museums. The relevant properties will then be used as a basis for the questions in the survey. Resolution This property determines how accurately the system can track people. A system with a high resolution can track people individually, which might be a concern for privacy. Personal Identifiability Personally Identifiable Information is any kind of data that can be used to identify the monitored person. This is generally considered a privacy breach [7], but might be necessary for the purpose of the system. An example of this is a (camera) system that monitors theft and vandalism. In either case, personally identifiable information of innocent visitors must be protected.

Property and Accessibility There is a concern with who can access the data. Outsiders with bad intentions shouldn’t be able to access the data, or influence it. Also relevant is who owns the data: can museums sell the data for their own gain? Retention Time It’s relevant to know how long the data is stored. If the captured data is deleted after a few days it can be considired to not invade privacy as much as a system that stores data for a longer time.

4.

OPINIONS OF EXPERTS

In order to find out what museums and artists want out of a monitoring system, and how sensor networks are currently used, interviews were conducted with professionals in the art world. Hans van Penderen is the head of the Technical Department of the Rijksmuseum Twenthe, a museum of fine arts and modern art [www.rijksmuseumtwenthe.nl]. The Rijksmuseum Twenthe is currently using sensor networks for security: if a sensor detects movement near an art piece, a camera is switched on. There are currently no other uses for sensor networks at that museum. The museum follows Dutch law with regards to privacy. Camera images are deleted after a certain time (he couldn’t tell me the exact number). There’s also a notice warning visitors about the camera surveillance. Any new system in the museum should follow the same rules [see also section The law ]. He was not very interested in analyzing visitor movements using sensor networks. Interactive installations could be used in the future. Hans expects that these will be purchased as part of an art collection, not actively deployed by the museum. It all depends on if it fits in the exhibition. 100% Fat is a media company that makes interactive pieces using sensor networks, for both museums and commercial companies. There’s interest for these installations from almost all types of museums. Museums are interested because the pieces 100% fat makes attract visitors. It’s all about presenting information to visitors in a novel and interesting way. 100% Fat does consider the privacy of visitors: installations are made in such a way that camera images are never saved. Jan Mensen is an art student at the AKI Media Art department. For his final exam he created a room with several media-based pieces. He is interested in a monitoring system, because it allows him to see which piece has the

most impact, and how people move between the pieces. He talked with fellow students at the AKI, he thinks that the things artists want from sensor networs is influenced by the type of art they make. It’s important for all types of art to see how people react to it. The artists usually has an hypothesis of a reaction from the crowd, it’s interesting for them to see if that expectation is correct. Jan would like to know how long people watch a piece. He wants to know if a different setup or lighting affects this. It’s also important to know how long people can stay to watch a video. A fellow student would like to know the emotional reaction, and the age of the viewers. There is some interest for using sensor networks in art pieces too, especially for Microsoft’s Kinect system. Sensors are already used (for media art) as a tool, for example to switch on a moving art piece if motion is detected.

5.

MONITORING SYSTEM: A PROOF OF CONCEPT

This research is not about the development of a fully functional sensor system. However, for the sake of the survey it’s important to have visible sensors that might have an effect on the visitors. It is not necessary that the monitoring system is functional at all: even in a real-life fully functional system, the visitors will never see the inner workings and the data.

Bill van Steenhoven is an artist who works with sensor networks. He expects that most art museums will not be really interested in an advanced sensor network, because they’d rather spend their money on a new art piece. There might be more use for sensor networks in museums that focus on (natural) history. These museums could use novel interactive installations to attract school classes and other visitors. Paul Klomp is an art teacher at the AKI Institute of Art and Innovation. According to him, a visible monitoring system will be seen as a piece of art itself. If it’s set up at a conventional museum, such a system will be conceived as a statement about ethics. His advice is to conduct the experiment in a modern art, or more precise media art setting.

4.1

The law

A monitoring system using cameras has to follow the privacy laws of the country it is in. Dutch law [10] requires that a privacy test is conducted. The utility of the system has to outweigh the interests and privacy of the people that are monitored. The employees or visitors have to be warned about the presence of the system, and hidden cameras aren’t allowed.

Figure 1. One of the dummy cameras that were used during the survey The ’system’ consists of four dummy cameras, one in each corner of the setup. The dummy cameras are replicas of actual security cameras, and contain a motion sensor so they can blink a light and rotate if they notice someone. This attracts attention to the camera, which is good for this purpose.

Note that this law might not apply to an interactive installation using cameras, as long as no images are saved. Also note that laws in other countries might be completely different.

4.2

Conclusions

It was often difficult to find the right person, since most people in this field don’t know much about sensor networks. In general, most people in the art world didn’t think about using sensor networks before the interview. Some museums like the Rijksmuseum Twenthe are using sensors for security purposes. These have to be simple and robust, and there isn’t really the need for something more high-tech.

Figure 2. The exhibition space where the survey was held

Artists like Paul Klomp and Bill van Steenhoven, and some art students, are creating interactive art using sensors. There’s a future in the field between tool and art. 100% Fat makes interactive media installations for museums, both the Twentse Welle and Rijksmuseum Twenthe are interested in these kind of installations.

The dummy cameras are positioned in each corner of a section of the exhibition space. The exhibition area of the University of Twente Creative Technology department at the Twente Bi¨ennale was used. Visitors could enter from two sides, see also figure 2.

The interviewed museums didn’t really have an interest for analyzing visitors, however the artists did. If such a system is made, it’s a good idea to give the information directly to the artists.

6.

SURVEY

A survey was held among visitors of a modern art exhibition, the Twente Bi¨ennale in Hengelo. This exhibition

focused on contemporary art by both international artists and local talent. The theme of this edition was the actuality of today’s society after the September 11 attack, and today’s culture of mass media (see also their website [13]).

6.1

Demographic

The people interviewed were all visitors of an exhibition on modern art. They can be described as being interested in art, and are more alternative and avant-garde compared to the average person. Among the interviewed were slightly more women than men, and slightly more young people, around 20 years old, than older people.

6.2

Survey Questions

The questions are based on the properties of monitoring systems that were described before. Most properties have one ’positive’ and one ’negative’ survey proposition associated with them. There are also some general questions about the use of these systems in an art setting. There was no question about retention time, instead a retention time of one week was given as context on the question form. The questions appeared in no particular order. During the survey, visitors were actively approached. They were told about the camera system, after which the cameras were pointed out to them. There were three possible answers to the questions: Strongly Disagree (1), Disagree (2), Neutral (3), Agree (4), Strongly Agree (5). 1. I would go to the museum less often if an advanced monitoring system is present 2. It’s interesting to use the sensor data for interactive art. 3. I feel uncomfortable seeing camera’s all the time. 4. It’s important that outsiders can’t hack into the computer system running the system. 5. The data from the system can be made public (anonymously) for academic research. 6. It’s a good idea to provide artists with the camera images, so they can see how the public reacts to their work. 7. The system can’t recognize me personally, that would invade my privacy. 8. I don’t mind being followed by a sensor network, as long as I’m told beforehand 9. The camera images can’t be used for purposes other than security. 10. The system data can’t be sold to commercial parties. 11. For some (security) applications it is necessary that visitors are recognizable. In these cases it’s not a problem that the privacy is invaded. 12. If the system recognizes groups and patterns instead of single people, it doesn’t invade privacy as much.

6.3

Survey Results

47 people have answered the survey. Some people did not answer all the questions. If a question was left unanswered, it was marked as neutral (3). 22 people responded that they didn’t like having cameras pointed at them. However, 27 people still wanted to know when they were being monitored. 26 people agreed

Figure 3. The results of the survey (N=47)

that resolution affects privacy, but only 3 people agreed strongly, and 14 people were neutral. A large majority of 31 people considered being recognized personally a breach of their privacy. Yet on the other hand 21 people agreed that this privacy breach was not a problem if the application demanded it. 26 people did not want the monitoring system to be used for other purposes than security, this seems problematic for a system tracking visitor movements. Only 3 people had a problem with using sensors for interactive art though. Perhaps the public does not consider an interactive installation to be a monitoring system, they would consider it to be a piece of art instead. There was a dilemma in the answers to questions 5 and 10. In the case of academic research, 35 people had no trouble with sharing their information. However, for selling to commercial parties it was the opposite: 40 people agreed that this should not be done, with 27 of them strongly agreeing. There are several explanations for this. Perhaps visitors on modern art exhibitions dislike commercial parties, visitors might even feel robbed if other parties make money out of their information. Perhaps it had to do with the way the questions were formulated. 38 people, almost everyone, considered it important that the system should not be broken into, this is not really a surprise. In the end, 40 people (85%, including neutral) replied that the presence of a monitoring system would not have any negative impact on their decision to go to a museum. The views of visitors did not really change

if they knew they were monitored. Some visitors quickly found out by looking at the cameras that they were dummies. This did not seem to impact their answers on the questionnaire. Most visitors either already had a strong opinion about privacy, or did not care much about privacy.

7.

[3]

CONCLUSION AND RECOMMENDATIONS

To conclude this research, the research questions are once again considered. What are the practical uses of monitoring visitors at an exhibition? Three possible practical uses for monitoring systems were proposed: A system that can track and analyze visitor movements, an advanced security system, and interactive installations using sensors. Museums were not interested in a more advanced security system. Privacy is not the issue here, they feel the current simple security systems suffice. There was some interest from artists for tracking visitor movements, but museums did not have much interest. Of the three proposed practical uses of sensor networks the interactive art installation has the most potential. It has real value for museum owners, because they will attract visitors. The visitors themselves mentioned they find these interactive art pieces interesting. They get something real in exchange for being monitored: an exciting experience. The monitoring system is presented in an attractive, artistic way. It is therefore expected that visitors do not experience the same discomfort as when they have ’ugly’ cameras pointed at them. Unlike a security system, the system can be designed to not store any personal data.

[4]

[5]

[6]

[7]

[8]

What do visitors consider ethical when they are being monitored? The conclusion from the survey is that most people do not care about privacy as much in the museum context. A sensor network can collect almost any data, as long as the visitor gets an exciting experience in return. An exception is in the case of personally identifiable information, the majority of people does not want to be recognized.

[9]

Can the type of exhibition influence a person’s views on privacy? The survey was held at an exhibition for modern art. It is possible that visitors in other types of museums have a different opinion on this subject. Further research is necessary to find out of the type of exhibition really has impact on the opinion visitors have on privacy. The same survey can be held on for example a museum for fine arts, or a museum for (natural) history.

[10]

Do visitors’ view on privacy change if they know they have been monitored? In general the interviewed visitors weren’t influenced much by the setup of dummy cameras. although more research is necessary to conclude anything.

8.

REFERENCES

[1] Kirsten Boehner and Phoebe Sengers. Opening the frame of the art museum: Technology between art and tool. In In Digital Arts and Culture (DAC, pages 123–132. ACM Press, 2005. [2] Lina M. Pestana Le˜ ao de Brito, Laura M. Rodr´ıguez Peralta, Filipe E. Sousa Santos, and Roberto P. Ramos Fernandes. Wireless sensor networks applied to museums’ environmental monitoring. In Proceedings of the 2008 The Fourth International

[11]

[12]

[13]

Conference on Wireless and Mobile Communications, pages 364–369, Washington, DC, USA, 2008. IEEE Computer Society. Bernhard Debatin, Jennette P. Lovejoy, Ann-Kathrin Horn, and Brittany N. Hughes. Facebook and online privacy: Attitudes, behaviors, and unintended consequences. Journal of Computer-Mediated Communication, 15(1):83–108, 2009. Su-Chu Hsu, Jin-Yao Lin, Carven Chen, Ying-Chung Chen, Jiun-Shian Lin, and Keng-Hau Chang. One million heartbeats. In Proceedings of the 15th international conference on Multimedia, MULTIMEDIA ’07, pages 365–366, New York, NY, USA, 2007. ACM. IkEnschede.nl. Alles is mogelijk! http://www.ikenschede.nl/opinie/1102-alles-ismogelijk-.html, April 2011. Katina Michael, Andrew McNamee, and MG Michael. The emerging ethics of humancentric gps tracking and monitoring. In Proceedings of the International Conference on Mobile Business, pages 34–, Washington, DC, USA, 2006. IEEE Computer Society. Arvind Narayanan and Vitaly Shmatikov. Myths and fallacies of ”personally identifiable information”. Commun. ACM, 53:24–26, June 2010. Edward J. Ottensmeyer and Mark A. Heroux. Ethics, public policy, and managing advanced technologies: The case of electronic surveillance. Journal of Business Ethics, 10:519–526, 1991. 10.1007/BF00383350. Sameer Pai, Philip Kuryloski, Henry Yip, Srikant Yennamandra, Stephen Wicker, Kirsten Boehner, and Geri Gay. Networks of sensors in public spaces: Combining technology with art. In Proceedings of the 21st International Conference on Advanced Information Networking and Applications Workshops - Volume 02, AINAW ’07, pages 396–402, Washington, DC, USA, 2007. IEEE Computer Society. College Bescherming Persoonsgegevens. Camera’s op de werkplek. http://www.cbpweb.nl/Pages/inf va camera werkplek.aspx, February 2011. Neil C Rowe. Interpreting suspicious and coordinated behavior in a sensor field. In Proceedings of the Meeting of the Military Sensing Symposium Specialty Group on Battlespace Acoustic and Seismic Sensing, Magnetic and Electric Field Sensors, August 2008. Flavia Sparacino. The museum wearable: real-time sensor-driven understanding of visitors’ interests for personalized visually-augmented museum experiences. In In: Proceedings of Museums and the Web (MW2002, pages 17–20, 2002. TwenteBiennale.nl. Concept. http://2011.twentebiennale.nl/index.php?option =com content&view=article&id=5&Itemid=5, May 2011.