Alexa nudges towards smart home energy efficiency

What unsustainable behavior needs to change:

Residential buildings are responsible for more than a quarter of total energy consumption in the US (EIA, 2020); and given that the majority of US and global residential properties still rely on fossil fuels, this takes a significant toll on the environment. While certain structural and mechanical measures have been taken to increase energy efficiency of residential structures, the occupants’ behaviour is also a great contributor to this high number. 

Two of the main barriers are occupants’ lack of awareness of their energy usage, along with a lack of information on how they could optimise it while maintaining personal comfort. Additionally, even if residents are aware and informed, they may not remember or have the opportunity to carry out the necessary actions. Considering the increasingly popular use of smart home assistants such as Amazon’s Alexa, and interconnected systems of smart home appliances; herein exists an opportunity for nudging towards optimising energy efficiency within homes.

Research into how smart home assistants can help reduce the environmental impact of residential energy use, via “bidirectional dialogue,” was conducted by a small team of engineering and communications academics from Virginia Tech and Florida State University respectively.

The Green Nudge:

The proactive smart home assistant aims to optimise energy efficiency by proactively making homeowners aware of their energy usage via: informative, conversational prompts, and autonomous actuation; that would reduce the effort required of homeowners to carry out the optimising actions (ie., Alexa would turn down the heating automatically, rather than the homeowner needing to do so). 

This was tested via an interactive online experiment with 307 US participants, who were provided a context-based scenario and an initial suggestion to increase energy efficiency. If participants’ responses were neutral or negative leaning to the initial suggestion (but not completely negatory), they were provided more information and tips, and their responses to this second suggestion were recorded as final. 

The result: a statistically significant increase by 16% of recommendation acceptances in response to the final suggestions compared to the initial suggestions. Conclusively, this outcome creates an opening for further “bidirectional human-building interaction” research towards sustainability.