Challenges Facing Ambient Computing
The potential of ambient computing is exciting and potentially life-changing for some, with many different benefits for users and society in general. However, there are six key challenges that must be addressed by the tech industry and beyond to make future ambient experiences a reality. These are:
Acceptance, availability, and affordability of new devices
Connectivity
Compatibility and interoperability
More compute and software capabilities in technology that is “built to last”
The increasing deployment of AI
Security and privacy
Some technologies that are needed for key ambient computing use cases are still not commonplace on the mass market or affordable for the general population, for example, devices like AR smartglasses or even hearables. Also, from previous experience, there is typically a prolonged period of cultural acceptance for these devices, as people get used to wearing and using them as part of their daily lives. This is supported by research showing that the perceived social stigma of using human hearing aids delays people using them for over ten years after their hearing deteriorates.
The wireless connectivity required to enable these complex future ambient experiences could be a challenge, with current connectivity still patchy among big crowds. The overarching challenge is the compatibility between the many different network providers and devices, which can slow down connectivity. To connect to different networks, devices have to go through different gateways and eventually the cloud for proper data exchange before going back to the device. This process must be improved to make ambient experiences quicker and more seamless.
Collaboration across devices from different vendors and operating systems is essential for ambient experiences, but far too complex today – even for relatively standard, straightforward devices that do not come from the same vendor. For devices to collaborate, they have to go to a cloud service that abstracts the local interoperability issues by centralizing signals. This slows down the overall experience and can also impact the safeguarding of security and privacy (more on that later).
One great technology example that captures this challenge is sensors. Sensors will undoubtedly be a big part of ambient computing in the future, providing the contextual awareness to gather information on the environments that people find themselves in, and then making these relevant at an individual level based on the person’s preferences. This requires greater compatibility between different sensors to ensure ambient experiences are ubiquitous, seamless, and relevant.
Ambient experiences require more compute power to support increasingly complex workloads, such as advanced AI for data gathering and learning. However, this compute power must take place across small sensors and other supporting technologies where low power and area are high priorities. Therefore, there must be a strong focus on efficient computing to meet these demands.
Furthermore, ambient compute systems must be built on highly capable hardware platforms that are “built to last,” especially if they are installed in areas that make them hard to replace once deployed, like the fabric of buildings. These hardware platforms also must be “software-defined,” with the ability to continuously add new, more advanced features via over-the-air software updates. Ambient experiences not only need a reliable, secure hardware foundation capable of managing large amounts of data, but also fluid software deployment so developers can keep adding more features and innovations.
As the volume of data increases across ambient compute systems, important and relevant information must be separated out from the rest of the noise. This processing of important, relevant information has to be compatible with the privacy, bandwidth, and energy efficiency requirements of ambient compute systems. AI and ML capabilities must be placed increasingly closer to the data sources to achieve this. This means more AI and ML compute workloads happening at the edge – on the actual devices – for quicker, more secure ambient experiences. Again, efficient performance is an increasing priority in making this a reality.
The collection of valuable personal data to create these ambient experiences means advanced privacy and security features are needed across all technologies. Despite many security-based technologies being available today, there must be a more widespread adoption of these existing technologies alongside additional measures and protections to appropriately handle sensitive information. This requires “privacy-preserving compute” to ensure the avalanche of data is protected, with secure maintenance and updates to devices also essential. The additional layer to this security challenge is the likely deployment of different levels of privacy for different situations and environments in public spaces.
