The study and its results come with certain limitations, which we discuss now.
First of all, whilst choosing a non-probability self-selection sampling method for this research allowed for quickly reaching participants that share traits with Early Adopters and allowed us to reach a sufficient sample size, it limits the generalizability of the study. The results should be taken as initial explanations for the low Use Behavior of IPAs and as initial recommendations for how to change IPAs rather than ultimate conclusions. Furthermore, the self-selection aspect of the questionnaire is likely to lead to people with a higher interest and awareness of the topic taking the survey, which could have led to people with more extreme positions participating (Zikmund et al. 2009, p.191), more moderate opinions could be underrepresented, potentially leading to overestimated effect sizes in the model. Concerning the size of the sample, as mentioned in chapter 3.2 already, the size of 97 participants is enough to uncover effects with a beta of more than 0.25 with a probability of 80% (Hair et al.
2014, p.21). Uncovering more subtle effects would require larger sample sizes. Yet as our research is focused on practical implications, uncovering only stronger effects is suitable since they represent the biggest opportunities for changing the target variable - the use of IPAs. Future studies could improve generalizability by using probability sampling methods with different tools for data gathering that have a lesser self-selection bias.
Second, it should also be noted that there are limitations to using a survey. It bears the potential of participants misunderstanding questions or not stating their true opinions or simply not recalling past behavior. Whilst we have tried to minimize the potential for these problems to arise through the use of tried and tested questionnaire items, multiple and rigorous pre-tests as well as removing items that do not satisfy the requirements for reliability, the potential for these problems remains (Zikmund et al. 2009, p.232). We suggest that future research adapts some of the UTAUT2 questions to be tailored to the field they are investigating more specifically and observes actual behavior.
Third, the fact that resources did not allow for a longitudinal study could have influenced the predicting variables captured through UTAUT 2, which are attitudes, and the target variable, Use Behavior. The study assumes that these attitudes towards IPAs, including the subject’s Behavioral Intention and Habit are more or less constant over at least one week and can therefore be used as a predictor for Use Behavior. A longitudinal study would be alleviated of this assumption.
Fourth, the use of UTAUT2 introduced some limitations to the study. While UTAUT2 gathers a fairly complete picture of factors influencing adoption, it comes at the cost of these factors being fairly theoretical, needing translation into practice. Future research could investigate certain aspects of the UTAUT2 model in more detail, trading breadth of the model for depth and bringing the results closer to practice. Additionally, the UTAUT2 model relies strongly on the assumption that behavioral intention induces adoption and does therefore not include as complete of a representation of the factors influencing behavior as a driver of adoption. As we established in 5.2, Habit is a stronger driver in the adoption of IPAs, so a more complete investigation of Habit might yield additional insights. We also suggest that the UTAUT2 model is extended by predictors for Habit. What is more is that the model does not allow the measurement of negative social influence. A closer investigation into whether there is a negative social influence in the case of Intelligent Personal Assistants could shed additional light on the barriers to adoption of IPAs.
Fifth, whilst this research aimed at discovering vectors of improvement for all IPAs, our research focused on phone-based IPAs as they are the most widespread and prevalent. Not all the results are transferable to other ways of delivering IPA services such as on the desktop or through speaker systems.
Sixth, most of our participants used the service in German. Since the technology is focused on language and the understanding of it, the English version might be better at understanding the user and therefore, the levels of variables like Performance Expectancy might change with the language. Yet, this is unlikely to influence the nature or the strengths of the relationship between Performance Expectancy and Habit but rather the levels of the variable itself.
Seventh, the self-selection bias in the data collected from Google Assistant users is likely to have distorted the data towards the very technologically well-versed since it is only available on cutting-edge smartphones. Additionally, the sample was very small. A further investigation on whether the Google Assistant has overcome barriers that the other IPAs have not could inform practice and advance the field.
Last but not least it should be noted that the field of Intelligent Personal Assistant moves very fast and that potential improvements this research highlights might already have been taken advantage of by the time it is published.
6 Conclusion
This study revealed first insights into the attitudes of users towards Intelligent Personal Assistants and these attitudes’ impact on actual Use Behavior, identifying potential drivers and barriers to the diffusion of the technology. Literature reviews on Artificial Intelligent in general and Intelligent Personal Assistants in particular, as well as on Adoption and Diffusion theory were conducted followed by the collection and subsequent analysis of data, allowing us to paint a thorough and holistic picture of the adoption process of IPAs for Millennials in Germany and Austria.
The results indicate that smartphone-based IPAs as a mean of interaction with the digital world lack a habitual anchor in the user’s life. According to our research, this is mostly due to the lack of utility generating use-cases, as well as the potential for a negative outside influence and heightened privacy concerns. Utility should be increased by improving the fit of commands to the interface-method, i.e. allowing for more specific commands, by allowing third-party integrations, clearly communicating functionality as well as improving natural language and voice recognition technology. Outside influence can be reduced by adding an input method less public than speaking out loud and care should be given to building
consumer trust on privacy concerns. This should be achieved whilst retaining the drivers to adoption - keeping the IPA easy to learn and fun.
Ultimately, it remains to be seen if users are going to find IPAs useful and are willing to adopt them by interacting with them on a regular basis. As our technologies get interconnected and smart home devices for control and entertainment are becoming a reality in our homes, there is the potential and need for an effective and simple mode of human-computer-interaction such as voice to become normal. The growing number of smart devices combined with the impressive progress in the field of voice recognition as well as AI-based information processing, will hopefully make for a time where interacting with them does not take up as much time and effort as they save us, freeing up cognitive capacity and increasing society’s global factor of productivity (Daugherty 2016).
7 Bibliography
9to5Google, 2016. Actions add 3rd-party features to Google Assistant, first partners launching soon for Home. 9to5Google. Available at:
https://9to5google.com/2016/12/08/google-assistant-actions-google-home/ [Accessed April 8, 2017].
Accenture, Future of Artificial Intelligence Economic Growth - Accenture. Available at:
https://www.accenture.com/us-en/insight-artificial-intelligence-future-growth [Accessed March 12, 2017].
Ajzen, I. & Fishbein, M., 2000. Attitudes and the Attitude-Behavior Relation: Reasoned and Automatic Processes. European Review of Social Psychology, 11(1), pp.1–33.
Apple a, iOS - Siri. Available at: http://www.apple.com/ios/siri/ [Accessed March 30, 2017a].
Apple b, SiriKit - Apple Developer. Available at: https://developer.apple.com/sirikit/
[Accessed April 5, 2017b].
Barclay, D., Higgins, C. & Thompson, R., 1995. The Partial Least Squares (pls) Approach to Casual Modeling: Personal Computer Adoption And Use as an Illustration,
Bhattacherjee, A. & Lin, C.-P., 2015. A unified model of IT continuance: three
complementary perspectives and crossover effects. European Journal of Information Systems, 24(4), pp.364–373.
Bolton, R.N. et al., 2013. Understanding Generation Y and their use of social media: a review and research agenda. Journal of Service Management, 24(3), pp.245–267.
Buhr, S., 2017. Barbie becomes a hologram version of herself. TechCrunch. Available at:
https://techcrunch.com/2017/02/17/barbie-becomes-a-hologram-version-of-herself/
[Accessed April 26, 2017].
Canbek, N.G. & Mutlu, M.E., 2016. On the track of Artificial Intelligence: Learning with Intelligent Personal Assistants. Journal of Human Sciences, 13(1), pp.592–601.
Chen, K.-Y. & Chang, M.-L., 2013. User acceptance of “near field communication ”mobile phone service: an investigation based on the “unified theory of acceptance and use of technology” model. The Service Industries Journal, 33(6), pp.609–623.
Choi, S., Kim, J. & Qu, M., 2015. The Flipside of Ubiquitous Connectivity by Smartphone-based Social Networking Service (SNS): Social Presence and Privacy Concern. In PACIS. pp. 1–13.
Cocotas, A., 2013. CHART OF THE DAY: iOS Users Are Wealthier, But That Doesn’t Explain The Monetization Gap. Business Insider. Available at:
http://www.businessinsider.com/chart-of-the-day-household-income-by-os-2013-3?IR=T [Accessed August 28, 2017].
Cohen, J., 1992. A power primer. Psychological bulletin, 112(1), pp.155–159.
Corbett, E. & Weber, A., 2016. What Can I Say?: Addressing User Experience Challenges of a Mobile Voice User Interface for Accessibility. In Proceedings of the 18th International Conference on Human-Computer Interaction with Mobile Devices and Services.
MobileHCI ’16. New York, NY, USA: ACM, pp. 72–82.
Creative Strategies, I., 2016. Voice Assistant Anyone? Yes please, but not in public! - Creative Strategies, Inc. Creative Strategies, Inc. Available at:
http://creativestrategies.com/voice-assistant-anyone-yes-please-but-not-in-public/
[Accessed March 12, 2017].
Cross, J., 2016. A list of all the Google Now voice commands. Greenbot. Available at:
http://www.greenbot.com/article/2359684/android/a-list-of-all-the-ok-google-voice-commands.html [Accessed September 1, 2017].
Daugherty, P., 2016. Artificial Intelligence May Change the Face of Business - Techonomy.
Techonomy. Available at: http://techonomy.com/2016/12/artificial-intelligence-may-change-the-face-of-business/ [Accessed March 12, 2017].
Davis, F.D., Jr, 1986. A technology acceptance model for empirically testing new end-user information systems: Theory and results. Massachusetts Institute of Technology.
Available at:
https://www.researchgate.net/profile/Fred_Davis2/publication/35465050_A_technology_
acceptance_model_for_empirically_testing_new_enduser_information_systems__theor y_and_results_/links/0c960519fbaddf3ba7000000.pdf.
Dee Dickerson, M. & Gentry, J.W., 1983. Characteristics of Adopters and Non-Adopters of Home Computers. The Journal of consumer research, 10(2), pp.225–235.
DroidLife, 2017. Distribution Archives | Droid Life. Droid Life. Available at: http://www.droid-life.com/tag/distribution/ [Accessed April 8, 2017].
Dwivedi, Y. et al., 2011. A Meta-analysis of the Unified Theory of Acceptance and Use of Technology (UTAUT). Governance and sustainability in information systems. Managing the transfer and diffusion of it, pp.155–170.
EF, EF EPI - Ranking der Englischkenntnisse nach Ländern. Available at:
http://www.ef.de/epi/ [Accessed June 18, 2017].
Enge, E., 2014. The Great Knowledge Box Showdown: Google Now vs. Siri vs. Cortana.
Stone Temple. Available at: https://www.stonetemple.com/great-knowledge-box-showdown/ [Accessed March 30, 2017].
Fenigstein, A., Scheier, M.F. & Buss, A.H., 1975. Public and private self-consciousness:
Assessment and theory. Journal of consulting and clinical psychology, 43(4), p.522.
Gallagher, S., 2016. The holographic anime “robot” that will keep house for lonely salarymen. Ars Technica. Available at:
https://arstechnica.com/information- technology/2016/12/the-anime-girlfriend-experience-gateboxs-ai-powered-holographic-home-robot/ [Accessed April 26, 2017].
Gao, Y., Li, H. & Luo, Y., 2015. An empirical study of wearable technology acceptance in healthcare. Industrial Management & Data Systems, 115(9), pp.1704–1723.
Hair, J.F., Jr. et al., 2014. A primer on partial least squares structural equation modeling (PLS-SEM), Sage Publications.
Hartmann, B. et al., 2005. Design and evaluation of expressive gesture synthesis for embodied conversational agents. In Proceedings of the fourth international joint
conference on Autonomous agents and multiagent systems. ACM, pp. 1095–1096.
Hauswald, J. et al., 2015. Sirius: An Open End-to-End Voice and Vision Personal Assistant and Its Implications for Future Warehouse Scale Computers. SIGPLAN Not., 50(4), pp.223–238.
Hill, I.D., 1983. “Natural language versus computer language” - Designing for Human-Computer Communication, London: Academic Press.
InternetLiveStats.com a, Internet Live Stats: Austria Internet Users. Available at:
http://www.internetlivestats.com/internet-users/austria/ [Accessed May 2, 2017a].
InternetLiveStats.com b, Internet Live Stats: Germany Internet Users. Available at:
http://www.internetlivestats.com/internet-users/germany/ [Accessed May 2, 2017b].
itim International a, Austria - Geert Hofstede. Available at: https://geert-hofstede.com/austria.html [Accessed August 21, 2017a].
itim International b, Germany - Geert Hofstede. Available at: https://geert-hofstede.com/germany.html [Accessed August 21, 2017b].
Joe Buzzanga, M., 2015. Beyond Keywords: The Revolution in Search. Available at:
http://www.sla.org/wp-content/uploads/2015/06/2015_Buzzanga.pdf.
Karsenty, L., 2002. Shifting the Design Philosophy of Spoken Natural Language Dialogue:
From Invisible to Transparent Systems. International Journal of Speech Technology, 5(2), pp.147–157.
Kharpal, A., 2016. Apple vs FBI: All you need to know. CNBC. Accessed September, 16, p.2016.
Knijnenburg, B. & Willemsen, M.C., 2014. Inferring Capabilities of Intelligent Agents.
Available at:
http://aisel.aisnet.org/cgi/viewcontent.cgi?article=1012&context=sighci2014.
Kondrat, J., 2017. Understanding asymmetries in ict acceptance at individual use in Poland:
the UTAUT2 theory combined with attitude, privacy risk, perceived security. Available at:
http://run.unl.pt/handle/10362/20004.
Kopp, S. et al., 2005. A Conversational Agent as Museum Guide – Design and Evaluation of a Real-World Application. In Intelligent Virtual Agents. International Workshop on
Intelligent Virtual Agents. Springer, Berlin, Heidelberg, pp. 329–343.
Lai, I.K.W. & Shi, G., 2015. The impact of privacy concerns on the intention for continued use of an integrated mobile instant messaging and social network platform. International Journal of Mobile Communications, 13(6), pp.641–669.
Lankton, N.K., Wilson, E.V. & Mao, E., 2010. Antecedents and determinants of information technology habit. Information & Management, 47(5), pp.300–307.
Laukkanen, T. & Pasanen, M., 2008. Mobile banking innovators and early adopters: How they differ from other online users? Journal of Financial Services Marketing, 13(2), pp.86–94.
Leswing, K., 2016. Here’s why people don't use Siri regularly, even though 98% of iPhone users have tried it. Business Insider Deutschland. Available at:
http://www.businessinsider.de/98-of-iphone-users-have-tried-siri-but-most-dont-use-it-regularly-2016-6?r=US&IR=T [Accessed April 5, 2017].
Luger, E. & Sellen, A., 2016. Like Having a Really Bad PA: The Gulf between User
Expectation and Experience of Conversational Agents. In Proceedings of the 2016 CHI Conference on Human Factors in Computing Systems. ACM, pp. 5286–5297.
Martins, C., Oliveira, T. & Popovič, A., 2014. Understanding the Internet banking adoption: A unified theory of acceptance and use of technology and perceived risk application.
International journal of information management, 34(1), pp.1–13.
Mennicken, S. et al., 2016. It’s like living with a friendly stranger: perceptions of personality traits in a smart home. In Proceedings of the 2016 ACM International Joint Conference on Pervasive and Ubiquitous Computing. ACM, pp. 120–131.
Munnukka, J., 2007. Characteristics of early adopters in mobile communications markets.
Marketing Intelligence & Planning, 25(7), pp.719–731.
Netmarketshare.com, 2015. Desktop Operating System Market Share. netmarketshare.
com/operating-systemmarket-share. Available at:
https://www.netmarketshare.com/operating-system-market-share.aspx?qprid=8&qpcustomd=1.
Novak, T.P., Hoffman, D.L. & Peralta, M., 1999. Building consumer trust in online
environments: the case for information privacy. Communications of the ACM, 42(4), pp.80–85.
Poggi, I. et al., 2005. Greta. A Believable Embodied Conversational Agent. In O. Stock & M.
Zancanaro, eds. Multimodal Intelligent Information Presentation. Text, Speech and Language Technology. Springer Netherlands, pp. 3–25.
Protalinski, E., 2017. Google’s speech recognition technology now has a 4.9% word error rate. VentureBeat. Available at: https://venturebeat.com/2017/05/17/googles-speech-recognition-technology-now-has-a-4-9-word-error-rate/ [Accessed September 8, 2017].
Purewal, S.J., 2017. The complete list of Siri commands. CNET. Available at:
https://www.cnet.com/how-to/the-complete-list-of-siri-commands/ [Accessed September 1, 2017].
Reynolds & Rodney, A., 2006. Handbook of Research on Electronic Surveys and Measurements, Idea Group Inc (IGI).
Rogers, E.M., 2010. Diffusion of Innovations, 4th Edition, Simon and Schuster.
Russell, S.J. & Norvig, P., 2009. Artificial Intelligence A Modern Approach Third Edition, Upper Saddle River, New Jersey 07458: Pearson Education Inc.
Saunders, M.N.K., 2011. Research methods for business students, 5/e, Pearson Education India.
Scheier, M.F. & Carver, C.S., 1985. The Self-Consciousness Scale: A Revised Version for Use with General Populations. Journal of applied social psychology, 15(8), pp.687–699.
Shukla Shubhendu, S. & Vijay, J., 2013. Applicability of Artificial Intelligence in Different Fields of Life. International Journal of Scientific Engineering and Research, 1(1), pp.28–
35.
Sterling, G., 2016. Google says 20 percent of mobile queries are voice searches. Search Engine Land. Available at: http://searchengineland.com/google-reveals-20-percent-queries-voice-queries-249917 [Accessed September 8, 2017].
Stift, C. & Yigen, B., 2017. Empirical Data.
Street-Porter, J., 2016. The rise of artificial intelligence risks making us all redundant. The Independent. Available at: http://www.independent.co.uk/voices/artificial-intelligence-ai-robots-mark-zuckerberg-unemployment-basic-income-minimum-wage-a7501806.html [Accessed March 12, 2017].
Taddicken, M., 2013. The “Privacy Paradox” in the Social Web: The Impact of Privacy Concerns, Individual Characteristics, and the Perceived Social Relevance on Different Forms of Self-Disclosure. Journal of computer-mediated communication: JCMC, 19(2), pp.248–273.
Theune, M., Heylen, D. & Nijholt, A., 2005. Generating Embodied Information Presentations.
In O. Stock & M. Zancanaro, eds. Multimodal Intelligent Information Presentation.
Springer Science & Business Media, pp. 47–65.
Turing, A.M., 1950. Computing Machinery and Intelligence. Mind; a quarterly review of psychology and philosophy, 59(236), pp.433–460.
Vardi, M.Y., 2012. Artificial Intelligence: Past and Future. Available at:
http://cacm.acm.org/magazines/2012/1/144824-artificial-intelligence-past-and-future/fulltext [Accessed March 12, 2017].
Venkatesh, V. et al., 2003. User Acceptance of Information Technology: Toward a Unified View. The Mississippi quarterly, 27(3), pp.425–478.
Venkatesh, V., Thong, J.Y.L. & Xu, X., 2012. Consumer Acceptance and Use of Information Technology: Extending the Unified Theory of Acceptance and Use of Technology. MIS Quarterly, 36(1), pp.157–178.
Yankelovich, N., 1996. How do users know what to say? Interactions. Available at:
http://dl.acm.org/citation.cfm?id=242500.
Yarow, J., 2013. CHART OF THE DAY: The Difference In Developer Revenue Between Android And iOS. Business Insider. Available at: http://www.businessinsider.com/chart-of-the-day-the-difference-in-developer-revenue-between-android-and-ios-2013-11?IR=T [Accessed August 28, 2017].
Zikmund, W.G. et al., 2009. Business Research Methods, Cengage Learning.