Blog Viewer

Artificial Intelligence and the Internet of Things (IoT)

By Gregory Ceton, CSI, CDT posted 12-11-2018 09:43 AM


Before I dive into the topic for this month, here’s a quick reminder to readers, if you haven’t read the Artificial Intelligence (AI) articles preceding this one in the series, you may want to catch up. The other articles aren’t very long, and they provide a good foundation for the concepts we’ll be covering this month. They are:

This month, we will be talking about another hot technology topic, the Internet of Things (IoT), and what it means for the advancement and use of AI as well as the benefits AI and IoT together can provide for facility performance and, potentially, construction project delivery.


The brain and the body

Perhaps the best analogy for the relationship between AI and the IoT is to think of an AI as the brain and connected IoT devices as the body. The connected devices are how the AI brain perceives and acts in the real world.

For purposes of this article, IoT can be defined as follows:

“The IoT is non-computer or mobile telephony devices with embedded software, electronics, and methods of connecting to the Internet to exchange data.”

IoT devices comprise the next step of expanding the Internet’s use to allow two-way communication by tools that have not been part of that exchange to date.


How big and how much?

The scope and financial scale of these developments are matched only by the rate of adoption and implementation. Current estimates say the IoT is made up of somewhere between 10 and 30 billion connected devices and is on pace to increase twofold by 2021 and double again before 2025.

Financially, the market for IoT devices and data is currently estimated to be around $200 billion. This is expected to go beyond $450 billion by 2020. Most of this growth is expected to take place in the manufacturing, transportation, and utilities markets, with additional significant development expected in the industrial sector. In most studies of IoT implementation, the construction industry has been largely overlooked, perhaps due to the sector’s off and on relationship with data and technological advancements. Just as we can expect to see increased implementation of AI in construction, we can look forward to the adoption and use of IoT devices in the building industry. Even if they are not employed directly in the project delivery process, familiarity with the growing capabilities and use of IoT devices will increase as owners plan to employ them in their facilities.


Unseen obstacles

The greatest obstacles currently facing IoT are the need for increased security and lack of dominant standards governing implementation and data exchange.

Similar to the challenges we faced with other applications in the construction industry, the data exchange obstacles have grown from a lack of industry consensus around an authoritative, small set of standards. A large catalog of IoT standards has been developed but they often compete with each other. It would be better to have a few, such as TCP/IP or HTTP, to define the work comprehensively. Without open standards, the result can be product lock-in and the potential for alienation from one’s own data sources, similar to what happened with building information modeling (BIM) and computer-aided facility management (CAFM) software. Since the codes are in flux, the data transmitted by IoT sensors and the communication protocols with Internet-enabled actuators and other controls lack the kind of standardization that much of the traffic on the web has had over its life cycle.

The construction industry has addressed security issues for its software and data tools in the same way the rest of the industries have, through increased focus on security protocols for traditional connected devices and data.

Since the software in IoT devices is “baked in,” it can be challenging to apply the same sort of protocols and standards to devices after they have been installed. IoT devices operate in the background, therefore evidence of security vulnerabilities may remain invisible. IoT sensors and actuators can be hijacked and added to a bot net capable of carrying out denial of service (DoS) attacks in the same way as any other Internet-connected device. In 2012, a researcher discovered a botnet called Aidra, which had compromised thousands of IoT devices and could easily be used for a DoS attack.


Where to from here?

It seems clear that despite these obstacles, the trend for increased connectivity will not subside. The most exciting aspect of all of this is the capacity to use the ever-growing, real-world data sets produced by IoT devices to train increasingly sophisticated AI and the impact the AI/IoT pairing can have on real world work and efficiency. Just like humans learn from their muscles and senses, it is possible that true general-purpose AI is unachievable without IoT. Expect to see more evidence of AI-enabled IoT devices in the construction and design arenas in the near future. These changes, much like the information-centric AI applications discussed last month, have the capacity to raise expectations about productivity and efficiency, an effect that may not be something you can choose not to engage with.

We’ll have a short column for the New Year in January. In early February, expect to read the last article in this series, in which we will look forward and try to draw some conclusions about what we can expect to see from AI next year, the next decade, and beyond.

Until then, happy holidays to all!


Gregory Ceton, CSI, CDT, is CSI director of strategic initiatives and special projects. He lives in the D.C. area and likes disruptive technology, good food, and cats.