Is Ambient Computing The Next Frontier Of AI?

Generative AI makes it possible, provided the right device is invented to make it work.

The concept of ambient computing envisions a future in which computing technologies will be seamlessly and ubiquitously integrated into our daily lives. It implies that humans will interact with those technologies in a natural and invisible way, without even thinking about it, as they will communicate intuitively with their owners according to their cognitive and sometimes emotional stimuli.

For a long time, this concept was associated with the development of Internet-connected devices meant to simplify communication between humans and their familiar equipment. The history of new technologies notably includes the initial experimental devices – Carnegie Mellon University’s Coca-Cola vending machine (1982), John Romkey’s Sunbeam Deluxe Automatic Radiant Control toaster (1990), the Trojan Room coffee machine of the University of Cambridge (1991) -, and then the first successful consumer products – LG’s smart refrigerator (2000), Fitbit Tracker (2009), Nest thermostat (2011), Ring’s Stick-Up Cam (2015), and of course, the Apple Watch (2015).

In this brief historical recap, I haven’t included two applications which, in my opinion, hold a unique place: Siri (2011) and Alexa (2014). Indeed, they demonstrated both the potential and the limits of artificial intelligence in bringing about ambient computing. Their potential lay in the fact that, for the first time, they functioned in a way that offered a glimpse of conversational human-machine communication, and in their integration into everyday devices (smartphones, speakers, headsets, computers, television sets…): The dual objective of natural and invisible interaction thus seemed achievable. However, Alexa and Siri also highlighted the limits of artificial intelligence at the time in understanding humans (use of coded phrases to interact with them, inaccessibility to their users’ environment) and the responses to their requests (inability to truly improvise).

Generative artificial intelligence undeniably has the potential to compensate for those two deficiencies. That’s why tomorrow, we will all have a individualized personal assistant, endowed with artificial intelligence, who will assist us in our personal and professional lives by being dedicated to our happiness and success and having access to all the data we will give it to achieve our goals. We will be able to ask it for factual information but also for advice. We will brainstorm with it but also share our feelings (excitement, frustration…) to better regulate them.

What remains to be addressed is the interface that will promote natural and invisible interactions with this assistant. In this regard, the smartphone has an obvious weakness: It is not always accessible because it is often tucked away in a clothing pocket or a bag, and for the same reason, it doesn’t have access to our environment. Moreover, a smartphone cannot (yet) be operated with voice commands only, and it’s not certain that all its applications could ever be. Using a smartphone is therefore an intentional act that doesn’t match the premises of ambient computing. This is why the vocal application of ChatGPT, despite its capabilities unmatched by Alexa and Siri, doesn’t fall into this category.

This is where the exciting current developments come into play. The most astonishing is the “Rewind Pendant,” a necklace that captures what its owner says and hears throughout the day, allowing the device to summarize it with the help of artificial intelligence. A necklace, unlike a smartphone, has access to its owner’s visual and auditory environment and thus resolves one of the problems mentioned above. But it creates another concerning the privacy of the people said owner meets or crosses during their daily activities. The founder of Rewind has responded to the many questions his device has generated since its launch in a video that is far from offering satisfactory solutions so that no one is recorded without their consent: If applied, they would call into question the interest of the system as it was initially presented. But the product deserves a chance, and we must recognize the merit of the innovators who take the risk of confronting the thorny questions presented by such unprecedented applications. Rewind Pendant is available for pre-order for $59 and with a subscription option. Initially, its interest could concern patients with memory disorders.

More interesting, even if they are not exempt from privacy issues, are the smart glasses recently introduced by Meta in collaboration with Ray-Ban. They are infinitely smarter than their predecessors (e.g. Google Glass introduced in 2013, Snap’s Spectacles launched in 2016) and offer a more natural use case than all the other “wearable” objects considered so far in the context of the continuity of use and access to the environment of their holder that are inherent to ambient computing. In this respect, they probably represent a better balance between performance and ergonomics than Apple’s Vision Pro headset: Glasses constitute a more natural and invisible interface, to take up the two criteria I set out in the opening remarks of this article, than an extended reality (XR) headset, which remains very visible and less easy to use continuously.

It is possible, even probable, that the equipment that will enable the advent of ambient computing and our next-generation personal assistants doesn’t yet exist: More than ten years passed between the emergence of the consumer Internet and the invention of the iPhone, which gave it its full power. In this regard, the explorations led by OpenAI, the inventor of ChatGPT, and Jony Ive, one of the inventors of the iPhone, promise to be captivating.