How AI and IoT must work together

This article was originally posted on VentureBeat.com

As artificial intelligence (AI) and Internet of Things (IoT) become more commonplace, it is both useful and important to understand how these two important trends work together to benefit specialists and the average person alike.

How can AI control and potentially improve the rollout of IoT?

For starters, IoT creates a tremendous amount of data. That data can be captured (enter: big data) and analyzed. But it is not realistic to do this tracking by human effort, given the sheer volume of data. AI will serve an important role in tracking — going through this mountain of IoT data and distilling it into actionable themes. Below are a few examples of how AI and IoT can be strongly synergistic in improving outcomes across areas such as power, health care, system design, and security.

Keeping the power on

For a good example of how AI can control and combine with IoT for impressive results, consider your home thermostat. Assuming it is connected and therefore an IoT device, what happens if it is an exceptionally hot day and the local power utility is experiencing brownouts? Traditionally, the system would overload, and then the staff of the utility would need to spend time and money to deal with angry customers to restore service.

In an IoT world, the utility staff can see how many devices are connected to the system, then react to avert a brownout by, for example, turning everyone’s thermostat up three degrees.

A built-in AI could alert the utility staff to an impending brownout that requires human action to avert. Or the AI could be more sophisticated and proactively turn thermostats up three degrees at homes and non-essential businesses, while keeping the thermostats stable at temperature-sensitive facilities such as hospitals and refrigerated warehouses.

The IoT system is a major advancement over the non-IoT system, but with the addition of AI, we have an even better outcome that can be prescriptive or proactive.

Making sure you stay alive

While staying cool is a luxury, staying alive is not. Though Fitbit watches and Apple Watches are nice to have (did you do your 10,000 steps today?), the widespread adoption of wearable sensors means that doctors and nurses can provide better health care than ever before. As our wearables become ever more sophisticated, instead of just reading your heart rate and pace when you run, they will be able to send an alert to your doctor if they detect that you might be having a heart attack. So IoT is a good start.

An AI control layer could take this to the next level. Imagine technology that receives that alert, then finds and routes the closest ambulance to you while notifying the hospital and your doctor. It then starts and drives your doctor’s autonomous car (an IoT device itself), delivering him or her to the hospital as quickly as possible. So adding AI to IoT could save the critical minutes to keep you alive. And guess what? A study published earlier this year shows that AI is already potentially even better than doctors at predicting that heart attack before it happens. Pretty cool, huh?

Boring but important stuff

AI will also be able to help IoT in two important under the hood areas: system design and security. The IoT includes everything from very basic sensors to sophisticated devices that have basic intelligence programmed in. At this stage, people make decisions on how smart to make each layer of IoT. As AI develops, though, it will translate the treasure trove of data coming out of IoT devices and identify where better sensors could provide additional insights that would help us make better decisions.

Conversely, AI can avoid overbuilding the system by understanding where additional data does not lead to better outcomes. In other words, we can save money by using more basic IoT sensors. And investing in more useful tools is good, right?

AI will also be able to help better secure the IoT world by anticipating and fighting intruders more quickly than human beings can. Machine learning is already starting to be used to predict safe IoT device behavior and help preserve cyber security. This trend will likely increase. Of course, those trying to penetrate IoT systems can also use AI to analyze weaknesses and exploit them — in many areas, AI is a double-edged sword.

AI and IoT can provide some pretty amazing benefits to people and the world if we use them sagely. One thing I can do that AI and IoT can’t do (yet) is remain optimistic, and hope that we continue to focus on improving our health, energy usage, and security in an ever-changing world.