Amazon, AI, and the Future of the Smart Home
Amazon and the Future of the Smart Home
Below is a short video I created exploring a prototype utilizing the technologies analyzed in the subsequent article (Published April 4, 2018).
This paper was written for the Digital Technologies module as part of my Master's in Digital Management at Hyper Island, Manchester (UK).
In this paper, the author will assess emerging technology trends and how they will disrupt home automation, commonly called the "smart home". The author will evaluate advancements in a suite of technologies in the Internet of Things as well as Artificial Intelligence. The company considered is Amazon, creator of the Amazon Echo, and a significant player in owning the control system for the smart home. In the end, the author will discuss some of the ethical issues presented by these technologies. Through video format, the author will create and test a prototype with an Amazon Echo stakeholder.
The Smart Home
Balta-Ozkan (2013) defines a smart home as "a residence equipped with a high-tech network, linking sensors and domestic devices, appliances, and features that can be remotely monitored, accessed or controlled, and provide services that respond to the needs of its inhabitants.” In 2015, Amazon became an early entrant into the smart home market with the debut of the Amazon Echo, a device capable of voice interaction with the purpose of controlling connected smart devices (Callahan 2015). However, Amazon faces obstacles in maintaining its share of the smart home market due to new competition from large tech organizations including Google, Apple, and Facebook.
The smart home has been considered a highly promising field for developing electronic technologies for many years, but certain key technical issues posed barriers to smart home development. This rest of this paper will focus on upcoming disruptions that will open new opportunities for smart home adoption, and provide Amazon with a competitive advantage in maintaining their favorable market position. They include: the Internet of Things (IoT) and Artificial Intelligence (AI).
The Internet of Things
The Internet of Things (IoT) is a network of physical objects that contain embedded technology to communicate and sense or interact with their environments. Gartner (2013) predicts that by 2020 the IoT will consist of up to 26 billion interconnected devices. Gateway devices, such as the Amazon Echo, will enable people, places, and things to participate in the IoT. Through interaction, an Echo can query an IoT device in its proximity and provide a bridge between technologies (Want 2015).
The network of physical objects that make up the IoT rely on sensory technologies which monitor and transmit information through wireless communication (Xu 2014). Early forms of sensor technology such as radio-frequency identification (RFID) served as the foundation of IoT. Advancements in camera technologies have created new high-content sensors, which provide rich sources of information both for human observation and for computer interpretation, and allow objects and devices to communicate and cooperate in new ways (Ding 2011).
Artificial Intelligence (AI)
To achieve its full potential, the IoT needs to be combined with a set of technologies categorized as artificial intelligence (AI). At its core, AI is about simulating intelligent behavior in machines of all kinds–and since IoT is about connecting those machines, there is a clear intersection between the two technologies (PwC 2017). As AI continues to evolve, the benefits for businesses will be transformational. As technologist Kevin Kelly (2014) states, "There is almost nothing we can think of that cannot be made new, different, or interesting by infusing it with some extra IQ.”
Neural networks are a form of artificial intelligence. They are computational models that mimic behavior of the human brain. While not nearly as complex as brains, they learn to solve problems efficiently through repeated exposure to data. They perform well at solving problems with many inputs, such as in facial or voice recognition. Creating a neural network requires may different processes to take place simultaneously. Until recently, this was a task a typical computer could not do, however decreased costs and accessibility of graphical processing units, or GPUs, now make this task achievable for common computers. In addition, deep learning, a computing method which optimizes the learning in neural networks, accelerated when they were ported to GPUs.
Neural Networks and Video-Based Sensors
Three-dimensional cameras can be aided by neural networks to “learn” tasks by considering examples of human behavior. Through computer vision, they can learn to identify images and evolve a set of relevant characteristics from learning the material that they process. Neural nets can be trained to recognize specific people, classify their activities, and respond to gestures, opening up a new space for gesture-based interaction models.
Current smart home technology is largely driven by voice technology, which is only useful in certain scenarios. Users often perceive voice-interaction as slow and embarrassing to use when other humans are around. However, neural nets and better cameras open up an entirely new space for gesture-based interaction. Gestural-interaction was popularized through devices such as the Wii Remote and Microsoft Kinect, but as it moves beyond the home gaming setting, freehand gestural interaction will enable users more convenient “walk up and use” interactions in everyday settings (Ren 2016).
In September 2017, Apple debuted the iPhoneX to the mass market, which included a camera that can perceive in three dimensions. These cameras create a point cloud that can read people in a scene, how they are posing, and how they are moving. With the advent of these new video-based technologies, IoT systems can be commanded using whatever body movements are easiest for the owner (Chan 2008).
Machine Learning & Big Data
Machine learning has been defined as "a subfield of computer science and artificial intelligence that deals with the construction and study of systems that can learn from data, rather than follow only explicitly programmed instructions.” (Jaffe 2014) AI requires thousands of examples of an object in order to categorize it. In an IoT situation, machine learning can take billions of data points and analyze it to find patterns that can be learned from. The realization of IoT depends on being able to gain the insights hidden in the vast and growing seas of data available. In order to efficiently scale, IoT is dependent on machine learning to find patterns that have the potential of enabling improvements in our daily lives (Jaffe 2014). Today the entire digital universe has become the teacher making AI smart (Kelly 2014).
Technological changes to the smart home come with several ethical considerations. The most prominent of these concerns are issues with access to and security of personal data, and the threat of several powerful companies with an unbalanced amount of power.
In order to tailor its systems to best support a user’s lifestyle, a smart home must collect information about the user. Camera-based sensors designed for IoT technology are made to monitor daily habits, which early adopters of the technology have perceived as “too intrusive, controlling restrictive, ‘big brother-like’ and engendering paranoia.” (Balta-Ozkan 2013). While organizations in the industry, including Amazon, have denied accessing consumer behavior and data without authorization, there is presently a lack of a clear policy on how it might be prevented if they choose to do so in the future (Barrett 2017). This will be a critical step in ensuring that data is kept confidential and that it is impossible for a third party to intercept data on purpose or by accident (Chan 2008).
The more people that use an artificial intelligence, the smarter it gets, rewarding consumers for tapping into the largest networks. As a result, Kelly (2014) notes that our AI future is “likely to be ruled by an oligarchy of two or three large, general-purpose cloud-based commercial intelligences.”
Improvements in the Internet of Things and Artificial Intelligence, allowing for new forms of communication and interaction with technology are poised to disrupt the smart home and companies competing to own the home automation hub.
The new technologies have the potential to bring benefits and create conveniences to consumers however there are significant barriers to overcome related to personal data and security, as well as the negative implications of advances of AI in a competitive marketplace.
Balta-Ozkan, N., Davidson, R., Bicket, M., Whitmarsh, L. (2013), “Social barriers to the adoption of smart homes.”, Energy Policy, September 2013.
Barrett, B. (2017). “Amazon’s ‘Echo Look’ Could Snoop A Lot More Than Just Your Clothes.”, Wired, April 2017. Available at: https://www.wired.com/2017/04/amazon-echo-look-privacy/ (Accessed 1st April 2018)
Button, K. (2017). “Smarter Collision Avoidance.” Aerospace America, June 2017, page 28.
Callahan, J. (2015). “Amazon Echo is now available for everyone to buy for $179.99, shipments start on July 14.” Android Central, June 2015. Available at: https://www.androidcentral.com/amazon-echo-now-available-everyone-buy-17999-shipments-start-july-14 (Accessed 2nd April 2018)
Chan, M., Estéve, D., Escriba, C., Campo, E. (2008). “A review of smart homes–Present state and future challenges.”, Computer Methods and Programs in Biomedicine, February 2008.
Ding, D., Cooper, R., Pasquina, P., Fici-Pasquina, L. (2011). “Sensor technology for smart homes.”, Maturitas, March 2011
Gartner, Inc. (2013). “Gartner Says the Internet of Things Installed Base Will Grow to 26 Billion Units By 2020.” Available: https://www.gartner.com/newsroom/id/2636073 (Accessed 2nd April 2018)
Jaffe, M. (2014). “IoT Won’t Work Without Artificial Intelligence.” Wired, November 2014. Available at: https://www.wired.com/insights/2014/11/iot-wont-work-without-artificial-intelligence/
Kelly, K. (2014). “The Three Breakthroughs That Have Finally Unleashed AI on the World.” Wired, October 2014. Available at: https://www.wired.com/2014/10/future-of-artificial-intelligence/ (Accessed 1st April 2018)
Kondori, F., Yousefi, S., Kouma, J., Liu, L., Li, H. (2015), “Direct hand pose estimation for immersive gestural interaction.”, Pattern Recognition Letters, April 2015.
PwC (2017), “Leveraging the upcoming disruptions from AI and IoT.” Available: https://www.pwc.com/gx/en/industries/communications/assets/pwc-ai-and-iot.pdf (Accessed 2nd April 2018)
Ren, G., Li, W., O’Neill, E. (2016). “Towards the design of effective freehand gestural interaction for interactive TV.”, Journal of Intelligent & Fuzzy Systems, 2016.
Xu, L., He, W., Li, S. (2014). “Internet of Things in Industries: A Survey.”, IEEE Transactions on Industrial Informatics, November 2014.
Want, R., Schilit, B., Jenson, S. (2015). “Enabling the Internet of Things.”, Computer, January 2015, Vol. 48 Issue 1, pp 28 - 35.
Yousefi, S., Kondori, F., Li, H. (2013). “Experiencing real 3D gestural interaction with mobile devices.”, Pattern Recognition Letters, February 2013.