Investigating Receptiveness to Sensing and Inference in the Home Using Sensor Proxies

1. Investigating Receptiveness to Sensing and Inference in the Home Using Sensor Proxies EunKyoungChoe, Sunny Consolvo, Jaeyeon Jung, Beverly Harrison, Shwetak N. Patel, Julie A. Kientz TsungYun 20130701

2. Outline • Introduction • Study Method • Analysis • Discussion • Conclusion

3. Introduction • Recent technical advances accelerate the integration of sensors into consumer devices in the home • Full-body 3D motion capture in games • Facial/voice recognition capabilities • Energy sensing systems • Wearable RFID in security systems But … is this a good thing? Privacy problem?

4. Introduction • Sensing and inference data captured in the home could be highly sensitive • Intimacy/secretive activities • Confidential conversations • Innocuous activities : cooking and eating [4] • Multiple stakeholders may have different perspectives on what is acceptable • Stakeholder: both householders and visitors

5. Introduction • This study investigates householders’ receptiveness to various sensing technologies • Offer a number of design insights which designers can use to reduce some concerns observed in the study

6. Study Method • Three phase • Initial in-lab session • Four weeks using sensor proxies in-situ • Exit interviews • Participants • 11 households (10 females, 12 males, aged 28-54) • various levels of education, occupation • owned a desktop/laptop with an average of 2.5 computers per household (min=1, max=5)

8. Study Method • (I) In-lab session • General population is not familiar with how sensing technologies work and what might be logged • Background survey and technology education session • Four sensing data: video, audio, electricity use, and movement

10. Study Method • (I) In-lab session • Encouraged participants to brainstorm possible application scenarios for each sensing technology • Consider the trade-offs (benefits/risks) • If participants were too positive about the sensors or applications, we probed about potential risks and vice versa

11. Study Method • (II) In-situ Phase • Cultural Probes method [7] • A technique used to inspire ideas in a design process • Probes : small packages that can include any sort of artifact (like a map, postcard, camera or diary) • participants record specific events, feelings or interactions using probes • Take-home packages

12. Study Method

13. Study Method • (II) In-situ Phase • 4 weeks • kitchen, master bedroom, family room, and child’s or guest bedroom/study room • not to turn on the sensor proxies during the first week • sensor light turned on whenever motion was detected

14. Study Method

15. Study Method • (III) Exit Interview • Ask participants about… • perceptions toward different sensing data • utility of the potential applications • issues regarding data access • issues regarding notification methods

16. Analysis • Perceived benefits and risks • Perceived benefits of in-home sensing applications • Perceived risks and concerns of in-home sensing • Tensions regarding sensing and inference • Tensions between couples • Tensions between parents and children • Tensions between householders and visitors

17. Analysis • Perceived benefits and risks • Perceived benefits of in-home sensing applications • Perceived risks and concerns of in-home sensing • Tensions regarding sensing and inference • Tensions between couples • Tensions between parents and children • Tensions between householders and visitors

18. Analysis • Perceived benefits of in-home sensing applications • People may be willing to accept invasive technologies if perceived benefits outweigh potential risks [19] • Applications directly related to household members’ health and safety • Monetary benefits and incentives

19. Analysis • Applications directly related to household members’ health and safety

20. Analysis • Still many participants were reluctant to the use of home automation systems

21. Analysis • Monetary benefits and incentives

22. Analysis • Perceived benefits and risks • Perceived benefits of in-home sensing applications • Perceived risks and concerns of in-home sensing • Tensions regarding sensing and inference • Tensions between couples • Tensions between parents and children • Tensions between householders and visitors

23. Analysis • Perceived risks and concerns of in-home sensing applications • Private Nature of the In-home Sensing and Inference Data • Unintended Consequences of Recording and Playback • Possibility of Data Leaks: Security and Data Storage

24. Analysis • Private Nature of the In-home Sensing and Inference Data

25. Analysis • Unintended Consequences of Recording and Playback

26. Analysis • Possibility of Data Leaks: Security and Data Storage

27. Analysis • Perceived benefits and risks • Perceived benefits of in-home sensing applications • Perceived risks and concerns of in-home sensing • Tensions regarding sensing and inference • Tensions between couples • Tensions between parents and children • Tensions between householders and visitors

28. Analysis • Tensions between couples

29. Analysis • Tensions between couples

30. Analysis • Perceived benefits and risks • Perceived benefits of in-home sensing applications • Perceived risks and concerns of in-home sensing • Tensions regarding sensing and inference • Tensions between couples • Tensions between parents and children • Tensions between householders and visitors

31. Analysis • Tensions between parents and children

32. Analysis • Tensions between parents and children • Not everyone agreed on including their child’s opinion in deciding whether to adopt sensing and inference systems

33. Analysis • Perceived benefits and risks • Perceived benefits of in-home sensing applications • Perceived risks and concerns of in-home sensing • Tensions regarding sensing and inference • Tensions between couples • Tensions between parents and children • Tensions between householders and visitors

34. Analysis • Tensions between householders and visitors

35. Analysis • Tensions between householders and visitors • Participants had different strategies for how they communicate an in-home sensing to a visitor • Depending on the relationship between the householder and visitor

36. Analysis • Tensions between householders and visitors

37. Discussion • Mechanisms to Reduce Privacy Risks • Limited capability sensors for the home • A new microphone-based cough sensor that only sends the relevant features of coughing sounds • Non-invertible audio processing techniques • Convert a general-purpose camera into a single event detector, e.g., fall detector

38. Discussion • Mechanisms to Reduce Privacy Risks • Context-aware sensing • Switch back and forth between high-fidelity (e.g., raw video) and low-fidelity (e.g., blurred video) sensing • High-fidelity when they are not at home and low-fidelity when they are at home • Automatically switch

39. Discussion • Mechanisms to Reduce Privacy Risks • Secure recording with limited playback • Enforce recorded data to be automatically deleted after a certain time period • To be viewed only a pre-specified number of times

40. Discussion • Tensions between Aesthetics and Visible Notification • Participants did not like to have sensing devices be visibly installed in their home • System sometimes needs to be hidden (e.g., supervising service people and babysitters) • Designing a gentle notification system (e.g., a location-based reminder on a cell phone) warrants future research efforts.

41. Conclusion • While in-home sensing and inference systems can provide numerous benefits, privacy risks and concerns exist • Conduct in-lab activities and four-weeks in situ with a cultural probe that used sensor proxies with 22 participants

42. Conclusion • Gather contextualized feedback on participants’ perceived benefits and risks of in-home sensing applications • Provide design insights to alleviate perceived privacy concerns and tensions

43. Q&A Thanks for your listening