IoT and AI are both technologies trending rapidly in the industry today. Both represent huge transitions into capabilities that only a few years ago existed as science-fiction. While both stand alone, their value is multiplied when combined. This combination represents a much better ROI for enterprises, organizations and governments who are sponsoring and deploying solutions. This symbiosis of technology will enable AI to be more meaningful and IoT to be more productive.
Why AI needs IoT
AI is comprised of many different mathematical algorithms to best answer questions across many different domains. Obviously, computers are good at math and now those same computers are so fast they can do the complex recursive calculations in real-time. That said, one thing AI always needs is training data.
For many companies doing AI today, that data has slowly been gathered over time with many users supplying the data. Whether it is people filling in the Google search bar, residents pushing buttons in elevators, or a receptionist in a medical clinic entering in patient information, large amounts of data have been collected in databases over the last several decades. This data is core to training computer models to correctly behave in an intelligent and predictive way. For years, this collected data has been a by-product of completing the core business function at hand.
IoT changes this process for everyone as devices and machines are far cheaper data collectors than humans. Sensors in a field, computers in automobiles, and smart tools all quietly gather and store data without any additional work on the part of the user. What previously required years of data collection and millions of dollars of investment is transformed into actionable insight by machines in a matter of months. This data, pumped into machine learning algorithms, and comprehended by experts, becomes the truly powerful AI we want in order to diagnose our diseases, drive our cars, and automate our industrial machines.
How IoT benefits from AI
At its core, IoT is nothing more than putting computers on all the things around us. Those computers usually don’t have a screen or keyboard, but instead do something basic like read a temperature, measure vibration, or turn on and off a light. IoT realizes its full potential when the data and the device are actionable. This means that not only is the data gathered from the sensors, but that the things in the ecosystem react in real-time.
The relationship of gathering data from one source and performing an action at another is sometimes called a feedback loop. For example, a temperature gauge sends the current reading from inside a milling machine, the milling machine has additional computers that turn the machine on and off. When the temperature reaches a certain threshold, the computer automatically turns down the RPMS of the motor and notifies the supporting technician of an emerging issue. logic to turn the motor off and on can be done via a very simple rule like “If the temperature is greater than 120 degrees, then turn off the motor.”
While many things in life can be optimized with simple rules, more often there is underlying complexity not accounted for. For example, maybe the motor should run hotter for short periods of time, or perhaps if the outside air is below 30 degrees, the engine should run to 140. These sorts of variables are less visible to the human mind, but are easily captured with AI. Taking the learned lessons of AI and bringing them into the decision process can completely optimize our processes. It means that the AI can identify what factors are truly relevant to the motor failing and what inputs have no effect. Essentially, AI can simulate the expert, who has gut intuition and years of experience about the systems we use. No longer does a specialist listen for a near inaudible noise, instead the AI systems can easily identify issues in the most optimal way. AI makes IoT actionable in the most powerful possibilities.
Putting them all together
It’s clear these two transformative technologies working in conjunction will enable businesses and cities to transform our everyday lives. Looking across the technology landscape, it will take technology, data scientists, domain experts, and system integration specialists working together to realize this vision. Hardware manufactures, connecting IoT platforms, cloud providers, and AI specialists will need to collaborate to create integrated architectures.