IoT and AI: The Anatomy of Our Digital World

In by contextere

This article was originally published on IoT Evolution.

Today’s digital world can be segmented into several transformative technologies that fall under the banner of the Internet of Things (IoT) and artificial intelligence (AI). This includes machine data analytics, automation, machine learning, computer vision, and natural language processing among others. Complementary technologies including augmented reality (AR) and wearables help users to access and interact with these capabilities. As each of these transformations begin to infiltrate our day-to-day lives, the question now becomes – how do they interact with each other and, more importantly, how do we interact with them?

From our perspective, one way to make sense of these transformations is to use a body-brain analogy. IoT can be viewed as the framework or body, made up of billions of connected devices. As with any body, there is a brain or mechanism that serves as the analysis and control hub for all operations. In the anatomy of our digital world, AI is this brain to interpret and control the IoT-based body. While that analogy works for a self-contained system, the role of humans in this system is critical to consider. AI, combined with human augmentation systems such as AR and wearables, enable human intelligence to be enhanced with increased contextual insights, knowledge, and awareness.

The Body

Though the proliferation of IoT devices is relatively recent, the concept of IoT has been around for nearly 20 years. The term was coined in 1999 to promote RFID technology; however, it didn’t truly catch wind until 10+ years later. So, what is IoT? Everyone tends to have their own definition, but at the core, it involves sensors within devices that are linked to communicate and transfer information with each other.

This transfer of information between devices is vital for three reasons: persistence, timeliness and relevance.

First, IoT enables machines to share critical information and creates a persistent data set that can be used for historical analytics.

Second, IoT enables a combination of cloud and edge computing to draw out insights from data that can inform decision making and automatically optimize device performance. Notably, the closer to the edge this computing occurs the faster actionable insights can be transmitted to the end user and the tighter the equipment performance optimization loop can be. In my opinion, this is significant since actionable data expires. The faster you get the insights to the end user, the more likely those insights will still be relevant.

Lastly, IoT data powers AI and, more specifically, machine learning (ML). ML, a subset of AI, “is concerned with conferring upon machines the ability to learn,” eliminating the need to program machines for every single action. ML is typically used to automatically optimize machine performance. However, we’re seeing more examples where ML can be used to provide contextual insights about machine performance and history to humans tasked with maintaining or operating those systems. To program or ‘train’ ML algorithms typically requires an enormous amount of data. “The more information there is to process, the more data the system is given, the more it learns and ultimately the more accurate it becomes.” This trend to increased relevance is one of the reasons why AI has exploded in recent years. IoT produces the data AI requires to be trained.

The Brain

Within the anatomy of our digital world, AI functions as the brain. In doing so, it absorbs the information from the IoT-enabled sensors/devices and determines appropriate recommendations to adapt and adjust system performance. The adjustments can be conducted automatically or with the support of human intervention.

Where humans are involved, AI provides another layer of engagement. Using the data to study and understand how humans think, learn and work to solve a problem, AI applies the key findings as a framework for intelligent recommendation engines and intelligent personal agents.

Through this process of technological transformation, humans have been successful in using AI to teach computers and robots to perform many human tasks. They’ve also enabled equipment to self-diagnose and self-optimize, effectively increasing the level of automation in the workplace. There is much talk of this automation displacing humans; however, while there may be disruption, I strongly believe AI will not replace humans.

Although AI can complete a large range of tasks, those tasks are still relatively specific. For example, AI can analyze large amounts of IoT data to predict when and how a machine will break and automate its control adjustments. AI cannot easily replace the hands-on activities that humans use to repair the machine. Instead, AI can be used to support the human in their process, enhancing the minute-to-minute decisions and judgment. AI and automation are unlikely to replace humans, but will instead replace specific tasks. In other words, it will be used to augment human productivity, freeing up their time to take on more complex and interesting responsibilities.

AI cannot easily replace the hands-on activities that humans use to repair the machine. Instead, AI can be used to support the human in their process, enhancing the minute-to-minute decisions and judgment.

The Bond

Together, IoT and AI combine to produce insights from data. They exist in a mutually-beneficial harmony, much like the body and mind. IoT creates the data required for AI, and AI provides a means to process the data to discover relationships and insights. The volume of data produced by IoT devices overwhelms human analysis; but, without that data, AI would not be as beneficial, as smart or as relevant. This relevance is crucial to humans in the workforce.

The ability of AI to deliver relevant and actionable insights to workers based on their context maximizes the value of IoT data. Humans play the role of executing on information and using today’s intelligent devices. In this view, IoT and AI become tools at our disposal. The interdependent and mutually-beneficial relationship between human, machine IoT data and AI will continue to empower each of us to do our jobs more effectively and safely.