There are concepts and technologies that come better as a pair, just like pen and paper, knowledge and power or nuts and bolts. Ok, better, the internet and routing, the blockchain and hashing, or digital twins, say a real time digital replica of a physical device.
IoT is about connecting machines and making use of the data generated from those machines, which is huge. In fact, IDC research group estimates that the amount of data created annually will reach 44 zettabytes in 2020 and up to 180 zettabytes (180 + 21 zeros) by 2025. And there is no end in sight to this flood of data as there are new cIonnected devices every minute.
This data needs to be processed before travelling through the networks to produce useful actions such as traffic control, climate prediction or crime detection. This is where AI needs to play an important role, be it by the means of Machine Learning, Cognitive Computing reasoning, natural language processing, speech recognition and vision (object recognition), human–computer interaction or dialog and narrative generation.
The point is stimulating intelligent behavior in machines. Could this be done some other way? Well, experts say that traditional methods of analyzing structured data are not designed to efficiently process the vast amounts of real-time information that stream from IoT devices. So, yes, quantity is an important factor in this equation.
The next level
That is the reason why the use of AI and Machine Learning is making a splash in Industrial IoT (IIoT) markets. Deloitte underlines that the combination of AI and IoT technology is helping companies “avoid unplanned downtime, increase operating efficiency, enable new products and services, and enhance risk management.” Major vendors of IoT platforms such as Amazon, GE, IBM, Microsoft, Oracle, PTC, and Salesforce have already learned some ropes and integrated AI capabilities.
Figures prove them right. IDC has estimated that in 2019, 40 percent of digital transformation initiatives will use AI services. By 2021, this figure will go up to 75 percent. On the other hand, Gartner says that more than 80 percent of IoT projects will include an AI component by 2022, in comparison to the current 10 percent.
Ricardo Santos, CEO of Heptasense and speaker at Internet Solutions World Congress (IoTSWC), acknowledges that adding AI to IoT brings the solutions to the next level. “IoT without intelligence is just big amounts of unstructured and meaningless data. AI has brought the tools for companies to understand how to leverage the value of all that information,” he says.
“Where Heptasense is concerned, AI enables the analysis of a remarkable amount of video in real-time and helps security teams detect threats without looking randomly at the cameras,” he adds.
There is no doubt that machine learning engines are a huge leap for IoT-based businesses where the ability to analyze, predict and automatically adjust to a particular need is highly prized. We’re not just talking about predictive maintenance, which is probably the brightest showcase of AI used in IIoT.
Artificial Intelligence is now being embedded in everything from logistics to healthcare, transportation, and agriculture and is expected to go much further in a variety of sectors. Platforms are there, but some experts are already arguing that putting AI at the edge, say within the devices themselves or on local servers rather than in the cloud, is the next goldmine. People constantly interacting with their digitally-assisted realities in real time will certainly require dynamic and competitive solutions.
Lasse Rouhiainen, author of Artificial Intelligence: 101 Things You Must Know Today About Our Future,holds that AI-powered devices are becoming smaller and able to perform more functions, with greater efficiency, behind the scenes. This is what he calls “Ambient Computing.”
“It’s highly likely that by 2025-2027, so many things in our daily lives will function in an ambient environment that it will be a bit the way electricity is today: something that is always working in the background, which we never think about until it stops working,” he says.
In this scenario, companies should begin to game out the potential impact of pervasive intelligence on their business, even if there are technical constraints, cultural obstacles, organizational barriers to adoption and other philosophical questions to be overcome.
These questions will be discussed at IoTSWC to help companies and organizations create their own roadmap in exploring AI’s new paths. For a reason: The advancement of AI is unstoppable. Considering this technology as a savior is woefully naïve. Yet predicting dystopian outcomes can cause potential solutions to be missed. “The new electricity,” as Andrew Ng described Artificial Intelligence and deep learning, deserves more.