The internet of things (IoT) is a complex expression of a simple idea: connecting a sensor to the internet in any time or place where it’s useful.
But there’s one formidable challenge: How do you make sense of all that sensor data? The best answer is to apply artificial intelligence/machine learning (AI/ML) algorithms that automate the collection, storage and analysis of IoT device data.
IoT devices come in countless forms — phones, wearables, smart speakers, video cameras, location beacons and things that haven’t been invented yet. In theory, IoT applications can converge data from all these sensors to generate unique insights and unforgettable customer experiences. In reality, AI/ML is mandatory to make everything work together.
A quick look at three emerging connectivity technologies of the 2020s — 5G networks, geospatial commerce and smart-cities applications — illustrates the crucial role of machine intelligence in IoT.
The next generation of wireless networks will generate a massive boost in mobile bandwidth in the 2020s. That will flood the internet with rivers of new data. IoT applications will inevitably spring up to leverage this new bandwidth.
Mobile devices and remote locations will provide two of the most logical use cases for 5G networks. It won’t just be consumers downloading 4K movies to their iPads. It’ll be medical providers live-streaming patient data to specialists in far-flung locales. Safety inspectors will use video and AI to scan for hazards that still images miss.
These high-bandwidth applications can generate mega-volumes of data that enable predictive AI. Algorithms that accurately forecast future outcomes allow people to be proactive vs. reactive, which has immense value. Now, imagine subtracting the predictive-AI component of 5G and IoT. The value proposition is far less clear.
Smartphone apps and location beacons enable the creation of geofenced areas that can document the positions of people and products in three dimensions in real time.
Geospatial Applications have profound potential in the commercial realm. Sensors can detect when consumers enter a specific section of a store, and notifications can direct them to the exact location of a product. On a cruise ship or in an amusement park, families can coordinate their activities and use mobile maps to find their favorite attractions.
But think about how much data these applications generate. And consider the ethical implications of tracking people’s behavior on such a broad scale. AI algorithms are essential to crafting consistent, highly personalized user experiences that stay within the bounds of humane privacy guidelines.
Smart cities represent the pinnacle of geospatial IoT applications. Sensors in traffic lights, bus stops and scooters will be able to work together to optimize traffic patterns and help urbanites travel to jobs, stores and entertainment venues.
In the public safety realm, IoT sensors and geofencing can help law enforcement agencies target their efforts where they can have the most impact. That could reduce the likelihood of unnecessary run-ins with people going about their business. Predictive AI could presumably make this kind of application even more effective.
While commercial IoT raises ethical concerns, smart-cities applications must deal with legal mandates. Again, this is an area where machine learning can help ensure that IoT initiatives conform with local ordinances.
The Power of AI, IoT and Analytics in the 2020s
5G networks and geospatial/smart-cities applications are only the most likely trends to emerge in the 2020s. More beguiling is the possibility that the application that pulls it all together does not even exist today.
Of course, a lot of uncertainties remain. Will 5G become pervasive enough to enable widespread IoT applications? Will public objections stall the progress of real-time behavior analysis? Will cities find the resources to embrace smart technologies?
We suspect the answers to these questions will emerge from use cases — people and companies trying, failing and trying again to profit from ubiquitous connectivity. The people who master the combination of AI, IoT and analytics cannot be known today, but it’s safe bet that they’ll be household names before the decade ends.
— Niraj Patel, director artificial intelligence