By continuing to use this site, you agree to our updated Privacy Policy and Terms of Use. ×








3 Prodigious Frontiers of Smart Device Evolution

3rd January, 2020 | Biometric Privacy | Entropic

Photo by Zdenek Machacek on Unsplash


In this article, we'll discuss the frontiers on which intelligent consumer devices are evolving. Specifically, we'll detail how corporations, including ones partially owned by governments are pushing the evolution of these devices with the objective of further gathering and curating information about individuals for the purposes of improving their products and services.

The resulting amassment and centralization of this information is a major cyberprivacy time bomb, that will totally eclipse the types of targeted device hacks that we are observing today.


What is Surveillance Resolution?

With an eye towards cyberprivacy, Entropic Labs regularly performs in-depth research on emerging intelligent consumer devices. This makes it easier for us to see the motivations behind, along with the patterns of device evolution.

Companies that develop and market intelligent devices to consumers to feed their growing addiction to the information of individuals, are pushing to maximize the quality, quantity, and diversity of captured customer information, with the goal of attaining the best possible insights into our way of life.

Within these frontiers are individual tracks of evolution that we'll detail. In some cases, individual tracks of evolution can fall under multiple frontiers.

Its important for us to comprehend this big picture, since it allows us to understand the frontiers that companies are pushing to get the information of individuals, what kind of information they are gathering and amassing, and ultimately who else might be motivated to hacking or acquiring the companies that have amassed this information.


To ease the explanation of this problem, we'll define "Surveillance Resolution" as the holistic surveillance potential that an organization has over an individual, that ultimately allows everything about them to be positively identified under varying conditions, effectively eliminating their autonomy and freedom.

While an organization usually has several channels that can be used to gather information about individuals, such as the basic use of their online services, the scope of our discussion will be on the smart devices that they produce. These devices empower them with a more 24x7 feed of their customer's lifestyle.

Targeting individuals or groups with nefarious intentions is likely not the objective of most organizations. However, their centralization of this diverse body of concentrated information, makes it a high value target to other entities, such as cybercriminals and state-owned acquiring companies. These entities may not be culturally predisposed to understanding privacy, and might be harboring more nefarious intentions. They can also be more motivated, and better equipped to refine this information to achieve their objectives.

This is by far the single biggest threat brought onto us by the Internet of Things.

The Evolving Frontiers of Surveillance Resolution

To further clarify the above diagram, we'll cover some examples within each main frontier of information gathering evolution.


QUALITY of information

Sensing Power: How are a device's onboard sensors evolving?

  1. Better Vision
    Companies that gather increasingly sharper still or video captures are better positioned to perform analysis, comparison, and recognition on the people (ex: facial profiles), and objects detected within the camera's field of view.

    Example 1
    Transitioning from using a camera with VGA to 1080P resolution in a device, allows higher resolution stills or video captures of an individual. This enables the ability to perform more accurate profiling of objects and individuals.

    Example 2
    Transitioning from using B/W to color night vision in a camera, allows color images of objects and individuals to be captured at night, dramatically increasing the amount of useable captured images for object and facial profiling.

  2. Better Audio
    Companies that gather sharper audio are better positioned to perform analysis, comparison, and recognition on the people (ex: voice prints), and other noises recorded within the device's listening range.

    Example
    Increasing the number of microphones in a far-field microphone array from 2 to 8, to empower a device with better listening capabilities in environments with more ambient noise.


QUANTITY of information

Staying Power: How is a device evolving towards the goal of sticking around and capturing as much information, for as long as possible?

  1. More 'ON' Time
    A device that can stay on longer is able to capture a higher quantity of audio information. Companies are evolving the ability of their devices to remain autonomously powered, without the need to be connected to a mains power source.

    Example 1
    Adding a rechargeable battery to a smart speaker, allows it to be powered when it is being used in alternate scenarios, such as outdoors and away from home.

    Example 2
    Transitioning a smart device from using a traditional plug-in power supply, to gleaning power provided through an existing USB connection.

  2. More Vision
    Companies that can capture more faces, or objects within their device's field of view, are better positioned to train facial and object recognition systems, using the people (ex: facial profiles), and objects detected within the camera's field of view.

    Example 1
    A version 1.0 smart display that initially recognizes only faces in its field of view can be improved in v2.0 to identify each individual that it captures, and respond differently based on the person identified.

    Example 2
    A version 1.0 smart video conferencing device, that initially has a regular/narrow angle camera, can be improved in v2.0 with a wide angle camera that allows it to capture more people and objects within it's field of view.

  3. More Audio
    Companies that can capture more audio within their device's listening range, are better positioned to train recognition of sounds from people (ex: voice prints), and other noises recorded within the device's listening range.

    Example 1
    A version 1.0 smart speaker that initially recognizes only human voices, can be improved in v2.0 to identify each individual voice print that it captures, and respond differently based on the person identified.

    Example 2
    A version 1.0 smart speaker that initially doesn't recognize ambient sounds, can be improved in v2.0 to identify specific types of sounds, such as the sound of glass breaking. Actions can then be taken, such as notifying the homeowner based on these events.

  4. More Comfort
    Companies that can make their smart devices as non-invasive as possible, stand to gather more information about individuals, because their devices can exist longer without being removed based on privacy or security concerns.

    Example 1
    Producing a smart speaker with subtle designs and colors, such as black and gray, can enable it to exist in a household in a "plugged-in" state longer, since people are less prone to noticing it.

    Example 2
    Producing a smart speaker with the ability to customize it with additional skins or covers, can help to increase its trust factor within a household.

    Example 3
    Producing a smart speaker with familiar cartoon-based form factors helps to increase the appeal of the device, especially to children.

    Example 4
    Adding a physical camera blocking switch to a smart display, can help to assuage privacy fears associated with the initial use of the device, with the goal of assuring the device's continued presence.

  5. More Convenience
    Companies that can make their smart devices more convenient to use can benefit from its increased usage, resulting in an increased volume of gathered information about its users.

    Example 1
    Evolving the voice assistant technology used by a smart speaker, to recognize additional contexts, phrases, languages, and accents.

    Example 2
    Releasing additional devices, such as a wifi mesh router that can interact with other existing intelligent devices in the household.


DIVERSITY of information

Outreach: How is the vendor evolving the ecosystem that uses their technologies, so as to capture as much diverse information as possible?

  1. Expand Form Factors
    While a company might initially develop intelligent devices as an initially closed ecosystem, they will commonly open up the technologies used in their devices for use by third parties, allowing integration into a broader set of alternate form factors, including ones that already have a trusted role. This increases the diversity of lifestyle telemetry they can gather about individuals.

    Example 1
    A company might initially produce their own smart speaker with voice assistant technology, that can be used to gather information from their customers based on specific use cases. Allowing a third party automobile manufacturer to integrate their voice assistant technology into cars, allows them to gather additional insights from the context of an individual's regular use of their car.

    Example 2
    A company that allows a third party appliance manufacturer to integrate their voice assistant technology into their refrigerators, allows them to gather additional insights from the context of an individual's regular use of their fridge.

    Example 3
    A company that produces a smart speaker might start producing security devices, such as video doorbells and window sensors, for the household that can be controlled by the smart speaker, empowering more household telemetry to be gathered.

  2. Expand Regions
    Expanding the regions in which they market intelligent devices allows an organization to open up their use to international markets. this increases the diversity of gathered cultural, language, and dialect-specific telemetry they can gather about individuals.

    Example
    A company might initially produce a smart screen to be marketed primarily in English-speaking countries, then later market the device to other more specific regions to refocus their gathered information, in advance of launching additional products and services to this region.

  3. Lower Cost
    Lowering the cost of their intelligent devices allows an organization to open up their use to more cost-prohibitive, or apprehensive markets. This increases the diversity of gathered cultural, language, and dialect-specific telemetry they can gather about individuals.

    Example
    A company that develops a v2.0 smart speaker that supercedes a previous v1.0 smart speaker might lower its cost, or even allow vendors to give it away with other purchases, to increase its appeal to more apprehensive markets.


How Can We Address These Problems?

The most important priority is for individuals to be empowered to quickly understand the implications of the intelligent devices they encounter in their daily lives.

Once we can understand these implications, we can be more mindful, and better equipped to make the right decisions to protect our privacy. With this knowledge, we can put the pressure back on the companies producing these devices to prevent the ongoing erosion of our freedom. With the overwhelming number of intelligent devices being produced by organizations with their burgeoning addiction to our personal information, this is a very challenging task.


  1. Know the Device's Capabilities
    If we really care about what a device is capable of over time, we need to focus less on what the companies privacy policy says, and more on the fundamental hardware capabilities of their devices.

  2. Know The Device's Manufacturer
    Know the device manufacturer, their investors and backing, and their government affiliations, as this is not static. It can change over time. Understand their track record of privacy violations, and human rights abuse, along with their government blacklist status.

  3. Stop Centralizing Personal Information
    Organizations that produce these devices must decentralize their customers stored personal information. This will reduce its concentration, making it less of a high value target, and demotivate its attack potential from nation-state actors, and cyber criminals.


Conclusion

Quantum computing will accelerate the processing, and infiltration of stored data amassed by the Internet of Things. Data processing tasks that are seemingly impossible now will become more trivial. This is discussed in further detail in this article originally posted by Vlad Andrei at Albaron Ventures.

The good news is that the same bleeding-edge technologies that can undermine our privacy, can also be used to defend our privacy. We'll discuss this in more detail in a future article.

If you have any feedback, questions, or suggestions, please let us know.


Acknowledgements:
Camera Lens (Modified) by Agence Olloweb on Unsplash
Closeup Photography of White Owl on Snow (Modified) by Zdenek Machacek on Unsplash