AI is shifting the IoT towards ambient intelligence

There is a notable shift across industries around how the traditional Internet of Things (IoT) is changing and evolving, with frontline data and data capture solutions delivering new value and intelligence with AI.

 

While some might call this the AI IoT, I think a better term is “ambient intelligence”. If the AI IoT is the “how”, then ambient intelligence is the “what” – the valuable outcome that frontline workers can leverage in their jobs.  

 

But what do we mean by ambient intelligence? Put simply, it’s digitising physical environments, workflows, assets and inventory, and turning that data into insight, trends, predictions, alerts, and recommendations for frontline workers and operations.

 

Ambient intelligence brings together the physical layer, the data layer, an AI analysis layer, and the execution layer with an appropriate level of automation. And it’s real-time, a living system or reflection of what’s happening in that moment.

 

Multi-sensor layers and AI can create an adaptive, real-time picture of operations by making physical spaces and frontline workers context-aware. This holistic approach connects the frontline, understands the whole environment, makes every asset visible, and supports intelligent automation in many industrial environments.

 

The connective tissue: distributed edge intelligence

A key enabler of ambient intelligence is the connective tissue of sensing technologies, software, edge real-time processing, and AI.

 

Physical environments feed AI models with diverse input data. This relies on integrating various sensing technologies to build a dynamic understanding of the environment and workflows, synthesised by modern software platforms.

 

This includes radio frequency identification (RFID), where advanced real-time location solutions (RTLS) use fixed and flexible systems to provide pinpoint locations. It can also incorporate computer vision, AI voice commands, device usage, geolocations, 2D and 3D machine vision, and barcodes for foundational identity data.

 

While multiplying sensors creates an exponential increase in image, sound, audio, performance and location data, multimodal AI models leverage this by ingesting these inputs to understand the environment and deliver multimodal intelligence and workflow automation.

 

Processing this data at the edge closes the gap between the physical environment, data, and execution, enabling zero-latency workflows required in high-velocity industrial environments and outdoor settings where connectivity may be irregular.

 

And with modern mobile computers and machine vision hardware engineered with neural processing units (NPUs), data is processed locally to deliver fast inference, reducing cloud costs and providing robust security and privacy. Frontline workers equipped with these devices receive actionable insights directly on screens or in the ear at the exact moment they are needed.

 

Looking ahead: augmented collective intelligence

While recognising that the IoT is evolving with AI into ambient intelligence, we should recognise that AI itself is developing. Computer vision, deep learning, and machine vision are now joined by AI agents or digital workers that will operate within ambient intelligent environments.

 

This combination of human workers and agents could be described as “augmented collective intelligence (ACI)”, with ambient intelligence being the fundamental fuel that agents operate within. This supercharges the frontline worker and delivers autonomous operations, which increases productivity and margins.

 

We can understand ACI as a distributed network of devices or a swarm of connected, dedicated agents / digital workers instead of a single all-knowing model. Secondly, ACI combines different styles of AI, such as generative and classical algorithms, to address complex tasks. And it requires human augmentation, where workers contribute unique intelligence and domain expertise to the network, while AI scales these talents through decision support.

 

At the technical level, ambient intelligence is about turning raw, multi-sensor data into real-time operational context through the combination of right-fit connectivity, multimodal sensors, and AI.

 

At an industry level, ambient intelligence supports a vision of autonomous, real-time frontline operations everywhere being digitised, automated, and intelligent.

 

And at the human level, it is about building environments where frontline workers are empowered with the foresight and clarity they need to make work better every day.

 

---

 

Stuart Hubbard is Senior Director, Artificial Intelligence at Zebra Technologies

 

Main image courtesy of iStockPhoto.com and alvarez