Recently a study was undertaken by IDC, sponsored by SAS, a global leader in data analytics software, looking at the coming together of AI and IoT. Together, AI and IoT are having a bigger than expected impact, leading to the term AIoT, the Artificial Intelligence of Things. In this article, we tap into the insights from that IDC-SAS study, and other sources, to look at how AIoT will affect us all.
The key finding in the IDC-SAS study is that combining AI with IoT in enterprise IoT deployments, significantly increases the value of the outcome. Why is this? The obvious answer is that IoT deployments generate huge volumes of data, and legacy tools simply aren't up to the task of interpreting it.
Shak Parran, quoted in the study, with Deloitte Canada references a mining company with 4,000 sensors which generate a terabyte of data each year. "There's really no way," he says, "for human beings to use all that data without AI and analytics".
Chetan Gadgil, Director of IoT at Intel, also quoted in the study, calls it "closing the loop". "Today most organisations are at the 'visibility' phase of IoT," he says. "where they can start to see what's going through IoT assets". Future phases will require significantly stronger AI capabilities.
The study shows that companies who deploy AI with IoT enjoy significantly better competitiveness than companies that deploy IoT alone. For example 32% of companies surveyed that had deployed IoT referenced improved productivity. That number increases to 52% when the IoT deployment was combined with AI.
Oliver Schabenberger, Chief Technical Offices with SAS, explains that "organisations working with IoT data realise that if they want to get the real value out of it, they need AI and analytics".
Iman Ghosh, writing recently in the blog site Visual Capitalist, calls the coming together of AI and IoT a "superpower of innovation". She highlights four markets where AIoT is likely to have the biggest impact: Wearables, Smart Home, Smart City and Smart Industry.
Gartner estimates that 80% of IoT projects will incorporate AI by 2022.
It's easy to think of AI as a tool that's used after data is collected in an IoT deployment, to apply analytics and to look for trends. But in fact what are called 'Edge AI' applications go much further, by not only analysing data at the IoT end point source, but by acting on it too.
In the Forbes article 'Why AIoT is Emerging as the Future of Industry' there's a great example. Think of a camera that is collecting images as part of an IoT application. It might send every frame captured, creating huge volumes of data for analysis. But Edge AI on the camera can ensure that images are only sent when certain objects are detected, or when certain events occur. Not only does this reduce data volumes and analytics load, but as a result it also improves response time when corrective action is required. Forbes points out that smart cameras are increasingly equipped with AI accelerators made by Intel, Nvidia or Qualcomm, to allow for edge AI applications of this type.
In fact Edge AI can be seen as one of the key technology drivers behind the growth of AIoT applications. As Michelle Duerst, Senior Director at Gartner points out, “Many executives do not understand how edge computing and AI can raise innovation capabilities, operational excellence and customer engagement. Industry leaders need to ask themselves how their product, service or customers can benefit from Edge AI implementation.”
According to Gartner, 91% of today’s data is processed in centralized data centers. But by 2022, about 74% of all data will need analysis and action on the edge.
Not only does Edge AI reduce data volumes and improve response times, but it also enables improved security. Edge AI applications can encrypt data at source. They can also ensure that only relevant data is sent, so that privacy is protected.
The coming together of IoT and AI, supported by Edge AI technology, creates a superpower that is likely to impact all of us in some way in the coming years.
REFERENCES