Thursday, December 19, 2019

Sensors in Homes

(Note: Ackama staff worked and volunteered on the original prototypes and systems for what is now Whare Hauora who are currently engaged in challenging Kainga Ora’s sensor procurement with a particular focus on privacy issues.)


In the last couple of days, there has been public argument on the Kāinga Ora project to put monitoring sensors in NZ state homes. Particular areas of concern being raised are about the nature of data being shared with government, issues of consent and vulnerability, and that this is individual level collection. By individual, I mean that it’s deep enough information to establish not just when people are home but also information on their behaviours. Additionally, people might not be given easy access to their own data without making a direct privacy request. 

Sensors in homes could be a powerful tool to impact the unhealthy state of our homes and understand the effects of our poor housing on New Zealand but they need to be done right and in a way that builds trust with everyone involved. This includes getting privacy and data sharing right. 

Sensors are surprisingly sensitive. If the people who find themselves cohabitating with them aren’t onboard with their presence they’re unlikely to give them the level of care necessary to get good data. Something as basic as putting a heater too near a sensor will completely throw off all its readings and that’s before taking into account anyone actually aiming to mislead the data (which we can be sure will occur to a few teenagers). 

Given this, it is deeply important that regardless of whether it’s Kainga Ora’s current project or a variation that goes ahead there needs to be a much clearer understanding of the broader benefits of sensors in homes and the processes that will lead to people accepting and gaining value from their deployment.

The benefits of sensors in homes ‘done right’

  • Health - both Whare Hauora and Kainga Ora are aligned in believing that improving health outcomes for New Zealanders is a key benefit of sensors in homes. Having good data like humidity and temperature in bedrooms drastically improves the ability of people to make changes that might improve their health (this is also a key area where making it easier for end-users to see their own data is likely to lead to them improving their own outcomes where possible, rather than waiting for government to respond).
  • Energy Cost - having sensors and power management in homes makes it easier to run and manage appliances at the right times. In addition, if users have access to their usage it becomes easier for efficiency ideas to actually help people reacting to the cost they bear. Current approaches of pre-payment lead to poorer households paying a much higher rate for their power and getting in a vicious cycle - some energy companies are innovating in real-time pricing, and there’s scope for power to be further subsidised along the lines of the winter power supplement. 
  • The Environment - as a follow on to better energy management, a lot of people don’t realise that a huge proportion of the high carbon power we still generate is targeted at supplying domestic power during the 6-7 pm time slot in winter (everyone gets home, turns on all their heaters, and starts cooking). Relatively minor reductions in power usage during this peak would significantly lessen our reliance on non-renewable power. Good sensors are a key part of allowing things like smart heaters to turn on at better or more efficient times so that people live comfortably while avoiding the worst of the peak.
  • Adoption of smart devices - this might seem like a surprising category in a discussion around sensors for some of the poorest in New Zealand, but prices for sensors and smart devices will continue to drop in future. The underlying cost of components for a manufacturer to add a wifi connection to most appliances is already less than a dollar, by the time any sensor networks being built now reach the end of their lives there will have been multiple years where wifi enablement has been the norm rather than a fancy add on.
  • Lowered mental strain - It’s well documented that one of the big challenges of poverty is the amount of time, energy, and worry that goes into thinking about and managing all the little challenges of daily life that become big challenges when every dollar matters. Automation and sensors can help make sure that you’re not spending money on heating when you don’t need to as well as lowering the mental burden on making that decision about when to turn the heater on and off based on actual health data.
  • Data - there are many uses for better data about what’s happening in New Zealand homes. Everything from holding landlords to account to meta-analysis of construction trends and weather effects will be interesting to someone. Although, it should be noted that there are very few potential reasons why this kind of interpretation should not be done on fully anonymised and aggregated data.  


How the Kainga Ora process could be improved

  • Explicitly focus on giving users a portal to see their own data - without access to their data most of the benefits of sensors are just not going to arise. It’s so much more efficient for someone to learn themselves that a room in their house might cause a health issue, and do something about it, than it is for there to be an entire cycle of engagement between government and individual. The eternal landlord refrain of “you just need to open some windows” is so much more compelling if you can immediately see data when you open the window, as well as being quickly disproven if the window does nothing.
  • Establish an arms-length governance committee - this committee should not be majority controlled by the government. The committee should be tasked with managing the process of aggregating and supplying data to appropriate organisations (including Kainga Ora). It should also be responsible for reviewing privacy concerns and policies. A diverse and representational governance committee will provide the needed insight and experience when reviewing these policies. 
  • Data should be shared appropriately with all relevant organisations - having gathered a data set it should be shared with community organisations and other appropriate parties as readily as it is shared with Kainga Ora. This allows other organisations to do work based on the data and findings without having to rely on a secondary process to generate public reports. Proactive sharing also makes it easier to use the data as a method of holding Kainga Ora and related parties to account.
  • Purchase portal and sensors separately - at present Kainga Ora is heading towards a bundled hardware and portal solution, this makes little sense as the sensor market is moving quickly and rapidly commoditising. The focus should be on building a great portal for both users and data aggregation that will support multiple devices and be open enough to integrate new devices that may crop up in the future. Sensors have already commoditised to the point that the number of devices being discussed could be procured directly as needed without breaking government procurement limits (this would also avoid the risk of piles of hardware sitting in a container somewhere while everything else is sorted out).
  • Don’t just get consent from people but show them some value - adoption of sensors shouldn’t just be something people accept, a real effort should be made for the user experience and data to be genuinely useful to whoever ends up with a sensor in their home. That includes treating any deployment as a marketing and education exercise on top of managing the actual physical installations.

None of these challenges are insurmountable and in the long run, the benefits of automation and sensor systems are absolutely there, but there needs to be careful thought and a focus on the users in whatever happens next in New Zealand. Technology is an amazing tool for making lives better but only if it’s done in a way that people won’t feel the need to “accidentally” put a small heat source next to it.

(This article originally posted on the Ackama blog)

No comments: