For years, we have played with the meaning of AI: many call it artificial intelligence, some say it is augmented intelligence or alternative intelligence. A few have stretched it further, comparing it with animal intelligence, especially the mysterious distributed intelligence of octopuses. But perhaps the most revealing phrase is an older one: ambient intelligence.
That term can be traced to the late 1990s, when Eli Zelkha and his colleagues at Palo
Alto Ventures used it in work linked to Philips Research. The vision was of technology so embedded in daily life that it would almost disappear into the environment like air or oxygen. It would be responsive, context-aware, adaptive and quietly present.
At the time, it sounded like a future of smart homes, connected appliances and ubiquitous sensors. Today, however, with generative AI, smart glasses, rings, cameras, microphones, wearables and agents, that old dream has returned with a more powerful brain.
A few weeks ago, at the Google I/O developer conference, ambient intelligence stopped being a research phrase and became a consumer product strategy. Google announced an AI-heavy overhaul of Search, deeper Gemini integration, information agents and intelligent eyewear through partnerships, including with Samsung, Warby Parker and Gentle Monster. Its new glasses promise directions, messaging, photos and Gemini-powered assistance. Search itself is changing from a box you type into, to a system that can plan, research and act.
Omniscient machine
Thus far, the internet waited for your intention. You opened a browser, typed a query, clicked a link, opened an app, made a choice. AI is shifting from intention to anticipation: it wants to see what you see, hear what you hear, remember useful context, understand your calendar, read your email, watch your screen, know your location and act before you have fully formed the question. This is what makes it ambient.
AI will not merely live inside a chatbot tab, but lurk in the spectacles on your face, the ring on your finger, the phone in your pocket, the camera at your door, the speaker in your living room, the dashboard in your car and the search bar through which you understand the world.
In some sense, we have been traipsing toward this for years. Nest tried to make the home responsive. Alexa and Google Assistant brought voice into the living room. Ring doorbells and home cameras turned neighbourhoods into semi-private surveillance grids.
Fitness trackers, smartwatches and health rings began collecting the body’s signals, as we aspired to become our “quantified self”. Humane’s AI Pin tried, unsuccessfully, to turn AI into a wearable companion. Many others are walking that path, and a successful one has been Meta’s Ray-Ban glasses where they smuggled intelligence into a familiar social object. The device looks like eyewear, but the function is ambient capture.
What generative AI changes is that all this ambient data becomes intelligible. Earlier, a microphone heard sound, a camera saw pixels, a watch counted steps. Now the model can use these signals to interpret the world around you. It can summarise a meeting, name an object, translate a sign, recognise a face, infer a mood, remember a preference and suggest an action. This ambient data of your surroundings becomes the fuel for intelligence.
AI and anxiety
This is both magical and unsettling.
Magical because ambient AI could be genuinely useful: A visually impaired person could navigate a street, an elderly parent could receive unobtrusive help, a student could have a tutor that understands the page in front of her, or a traveller in Tokyo could read signs and hear translations. Like Sanjaya in the Mahabharata, AI could become the seer who narrates the distant battlefield in real time.
But Sanjaya was a trusted retainer of Dhritarashtra. The same cannot be said of our ambient AI.
It will be built by Big Tech like Google or Meta, trained on our data, monetised through business models we barely understand, and governed by flimsy guardrails that substantially lag the technology.
The camera on someone’s glasses does not ask for permission before it sees a person nor does a meeting recorder always know who has consented. Amazon’s Ring cameras already raised questions about neighbourhood surveillance and non-consensual recording. Ambient AI will only magnify those anxieties.
This is why resistance to AI will intensify. In the early days of Google Glass, people wearing them were called “glassholes”, with some humour, and a lot of derision.
What happens to human agency?
We are already seeing resistance to data centres—their energy use, water consumption, land demands and environmental footprint. Those data centres are the visible temples of AI. So will be the symbols of ambient intelligence—the rings, glasses and pendants. People may accept AI that answers a question, but will resist AI that hovers around them.
It won’t be as much about privacy, but a deeper issue: of agency.
When an AI answers a prompt, the human remains the initiator. When ambient AI anticipates, nudges and recommends, the balance shifts to the technology. It shapes not only my attention, but my intention. So the fear will not merely be that AI will take away human faculties, but that it may quietly take away human agency. Ambient AI may begin to resemble a technological god.
Omnipresent, always around us; omniscient, remembering and knowing everything; and seemingly omnipotent, able to act across our digital lives. But gods inspire devotion and fear in equal measure and they demand rituals, sacrifices and surrender.
The smartphone demanded our attention; ambient AI might demand surrender.
Author is co-founder of AI&Beyond. Views are personal.
