Most of the focus in the media has been on smart electric meters and the many risks that these devices pose for customers and their families. Another risk relates to privacy invasions associated with smart water meters. A growing number of utilities are installing these devices.
The sales pitch from smart water meter vendors includes the following:
Water Conservation and Restrictions: When water conservation moves from being a priority to a necessity, smart water meters allow water utilities to react quickly to government mandates for water restrictions and to efficiently identify violators. It facilitates this process by providing utilities the ability to view customer usage on an hourly or sub-hourly basis, allowing them to detect excessive flow on “off” days.
Leak Detection: Utilities can minimize the problem of leaks on the customer side. Leak detection-enabled smart meters allow utilities to identify continuous water usage over specified periods of time. When continuous usage is detected, the system sends leak alerts so the utility can notify the customer about the potential leak.
The above advertised “features” for smart water meters means that you would be subjected to constant surveillance, even though no mention is made of customer privacy invasions by vendors or utilities. Moreover, combined with the granular data collected by smart electric meters, an even clearer picture emerges on your behaviors in the home.
Quoting a paper entitled, “Privacy Technology Options for Protecting and Processing Utility Readings,”  by computer security and privacy expert George Danezis, dated March 1, 2013:
“Even more intrusive information can be inferred when combining electricity with other utility readings, for example water and gas readings. Such combined readings can be used to detect different patterns of cooking in a household, since cooking activity exhibits correlated uses of electricity, gas and water. Similarly, the frequency of use of a dishwasher or washing machine can be inferred. Finally, the combined use of large volumes of water along with either gas or electricity can be used to infer how often members of the household have showers. Electricity and water provides information about night time patterns of sanitation, and even how often and when inhabitants use the toilet overnight.”
Research has already been conducted to develop algorithms to disaggregate water usage sampled at rates between 15 minutes and 3 hours. As stated in a paper entitled, “Activity Analysis Based on Low Sample Rate Smart Meters” :
“Activity analysis disaggregates utility consumption from smart meters into specific usage that associates with human activities. It can not only help residents better manage their consumption for sustainable lifestyle, but also allow utility managers to devise conservation programs. … We propose a novel statistical framework for disaggregation on coarse granular smart meter readings by modeling fixture characteristics, household behavior, and activity correlations. … Interesting activity-level consumption patterns have been identified, and the evaluation on both real and synthetic datasets has shown high accuracy on discovering washer and shower.”
From the paper presented at a conference in 2011 on Knowledge Discovery and Data Mining, the figure below illustrates water meter data versus expected disaggregated activities.
The results of the study showed high confidence in recognizing washer use and showering; toilet flush recognitions was still good at the sample rate of 15 minutes; at a sample rate of 1 hour, the algorithm was only about 50% accurate at identifying toilet flushes with the best algorithm.
As shown in the highlighted published article, 15-minute water meter interval data can be used to infer customer toilet flushes, showers, and use of a washing machine with fairly high confidence. The confidence levels drop significantly as the interval length of the data extends to one hour and beyond.
This article was written in order to raise the awareness of consumers to the privacy invasions that are associated with smart water meters that collect granular data.
Furthermore, consumers need to be aware of how combining smart water meter interval data with the data collected from the smart electric meters amplifies the degree of privacy invasions. In this sense, the total is greater than the sum of the parts. Possible ambiguity on deciphering the difference between using a washing machine or a dishwasher or more precisely determining when someone took a shower while using increased lighting and a bathroom fan would be eliminated with the added information of how many gallons of water were used at the same time as increased usage of electrical energy.
Even if some people think that smart electric meters or smart water meters each by themselves do not unreasonably invade a person’s privacy, where does it stop as additional so-called “smart” products are added to the mix? At some point we reach a tipping point where limits must be placed on what can be considered a “reasonable” invasion of privacy. To do this, we need more people to take note of the issues and speak up to ensure that the privacy interests of all consumers are seriously considered before new technologies are imposed upon them. Let’s make this happen before it is too late.
 “Privacy Technology Options for Protecting and Processing Utility Readings,” George Danezis, dated March 1, 2013; this document was prepared as background material for working groups dealing with privacy issues related to granular data collected utility smart meters. The entire document is available here: Privacy Options by G. Danezis 2013.
 “Activity Analysis Based on Low Sample Rate Smart Meters,” by Feng Chen et.al., Proceedings of the 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2011; 10.1145/2020408.2020450.