7 business lessons from AI at the edge
Blog: Capgemini CTO Blog
Zhiwei Jiang, CEO, Insights & Data and Ron Tolido, CTO, Insights & Data, Capgemini give you seven fascinating examples of Edge AI, including the business lessons each of them holds.
Data-powered enterprises use data, analytics, and AI to fulfill their corporate purpose, achieve their business objectives, and drive innovation. And in today’s shaky world, they need to be like water to get there: adaptive by design, hyper-agile, super-responsive. There’s a place they find themselves increasingly to build and use these water-like capabilities: At the very outskirts of the business, far from central governance and IT, close to clients, partners, employees, the shop, the factory, the truck. It’s out there in the field where up-to-date data needs to be collected in real-time, analyzed on the spot, and turned into split-second actions, without being dependent on a corporate backbone that simply is too slow and too far away.
Edge AI – the use of AI in physical devices and all sort of other ‘things’ – is a perfect enabler for these new dynamics: Equipped with sensors and microprocessors, even the tiniest of things can collect data, analyze it with powerful AI algorithms (such as deep learning) and turn it into immediate action. And all of that is done with specialized technology that is typically inexpensive and has a small footprint in terms of power consumption. Moreover, it doesn’t dependent on a continuous, high-bandwidth connection to a central network.
Not all organizations may yet bet their future on Edge AI, if only because there is a minimum number of physical assets involved in their business model. Even then, current developments in Edge AI hold key business lessons about how to become water-like ‘at the edge’ of the enterprise – think about autonomous decisions, real-time responsiveness, saving energy, infinite scalability, data security, privacy-compliance, and always-on availability.
Also, if nothing else, Edge AI use cases are simply compelling and inspiring. We have selected seven of our favorite examples to whet your appetite:
It seems Ingenuity stole the show, right? The small robotic helicopter made history with its first powered, controlled flight by an aircraft on another planet. Yet, it was brought to Mars by Perseverance. The rover managed to land safely with its unique Terrain Relative Navigation, using AI image recognition to select an non-hazardous landing site in real-time. Not something that you can manage centrally, as the communication latency between Mars and Earth is 20 minutes. It’s a powerful metaphor, showing how to achieve critical business responsiveness and adaptability without centralized control.
2. Elephant Collar
Elephants are an endangered species, illegally hunted for their ivory tusk and – even worse – just for fun. Edge AI is coming to the rescue to help preserve these gentle giants. Using a smart collar, various solutions are built to detect human presence, monitor elephant activity, and carry out other intelligent, protective activities. The secret inside the collar is called TinyML: AI machine learning capabilities on very lightweight processors running on minimal power for an extensive period. The lesson: putting key activities at the edge of business can considerably save energy, traffic and bandwidth, reducing carbon footprint and cost.
3. Electronic motorbike
Back then, there were no electronic heavy motorbikes in Norway. So, our colleague Robert Engels built one himself. Over the years, besides being fun to drive, it became a platform to test new, innovative hardware and software. Its latest addition: a built-in AI system (just above the taillight) that uses image recognition to detect approaching ambulances and police cars. Not to avoid speeding tickets, but to timely make way for emergency vehicles. Enterprises can take similar measures by detecting anomalies early – as and where they occur – rather than waiting for them to ripple through to a central authority.
In ancient times, people believed that oceans marked the edge of a flat world. Indeed, quite an appropriate place to pioneer with Edge AI. IBM’s Mayflower Autonomous Ship is cramped with sensors and intelligent devices. First, to navigate the vast ocean with no one onboard, avoid obstacles, learn how to reach the next harbor. But also to collect and analyze key environmental data, for example regarding the whale population, micro plastics, and ocean chemistry. The enterprise lesson: Breakthrough, innovative insights are for the grab night and day at the edge of business. And no hands on deck needed to get it done.
5. Drone Swarms
A drone is a fascinating machine. But how about lots of them, like a swarm? Take this swarm of autonomous drones equipped with LiDAR (Light Detection and Ranging) sensors to map a whole forest, providing crucial insights for forest preservation. The AI algorithms have been trained in virtual, simulated environments to properly identify trees, but also to avoid collisions. Edge AI enables the swarm to explore a forest autonomously, yet offloading its collected data to the cloud afterwards for analytical purposes. Key learning? Scaling at the edge is easier, faster and more versatile than trying to stretch a central facility to its limits.
6. Blue Eye
Whether we like it or not, we’ve all become quite familiar with the need for social distancing and wearing masks. In practice, it’s difficult to ensure compliance to the rules, as it requires substantial manpower and may intrude personal privacy and data security. What if the solution came as a box, equipped with Edge AI? Capgemini’s Blue Eye has specialized Intel technology inside, using AI to detect insufficient distancing and unmasked faces. As it is not connected to a central system, privacy and security are 100% guaranteed – which is reassuring. What happens at the edge, stays at the edge – quite a solid business bet.
7. Spaceborn Computer-2
Talking about the edge: let’s turn to space, the final frontier. The International Space Station (ISS) may be in a relatively low orbit, but the conditions are harsh, and Earth is far away. Now, powerful Edge AI has been rocketed onboard. HPE’s rugged Spaceborne Computer-2 ingests data from a range of edge devices, including satellites and cameras, and has the AI to support astronauts in conducting a wide variety of experiments self-sufficiently. It’s used to monitor the crew’s health as well. Business takeaway: equip the outer rim of the enterprise with the best tools and capabilities, so they can explore for the rest of us.
Edge AI is still evolving, and the technology may not already be suitable for every enterprise. But it is crystal clear what will drive the next wave of business innovation. It’s Simply the Edge. And when it’s unleashed, it’s better than all the rest.