The times during which artificial intelligence (AI) required cloud connectivity to function are quickly moving past us. The manufacturing of low-power-consuming chips that cater AI and machine learning (ML) features, along with the development of efficient AI and ML algorithms that can directly be run on them, has slowly pushed AI to independently function on devices. The inference system-on-chip, developed by MIT, can perform 3 to 7 times faster than previous chips while consuming 95% less power. Even Apple’s latest iPhones include an AI chip called the Neural Engine that performs machine-learning tasks such as image recognition on the device.
As a result of such chips being integrated into electro-mechanical devices, the resulting “smart” device ensures portability and showcases lower latency and connectivity independence. It’s estimated that by the end of 2018, around 500 million devices will carry chips that support AI. In addition, the annual shipments of such smart devices are expected to increase from 79 million in 2017 to 1.2 billion by 2023. As almost all large tech firms – including Intel, IBM, Microsoft, and Google – are involved in advancing AI technology, smart devices are set to become more commonplace and pervasive. Powered by devices that carry AI-enabled chips and microprocessors, we are now entering an era of “pervasive” AI. So, how would such smart devices upend several industries around the world?
In manufacturing, AI chips embedded into assembly line robots would make them more efficient at their jobs – calculating the motion of their arms a thousand times faster than they currently do. These robots would also develop quick reaction times to disruptions and work better with other robots and humans. In large warehouses, a system of such AI-empowered robots could coordinate with each other to efficiently pick, pack, and ship orders. It’s no wonder, that Amazon has already put such robots to work in their distribution centers, to aid its vast logistics and distribution problems.
The construction industry would also benefit as a result of AI-powered drones, robots, and connected cameras. With real-time monitoring of construction sites, productivity and efficiency would surge along with a reduction of waste materials. In one case, a project using Doxel’s progress-monitoring robot saw a 38% increase in productivity and an 11% reduction in estimated project cost. On the downside, efficient material management would reduce the demand for construction materials, which could garner losses for material providers. As windmills and wind farms become prevalent sources of generating renewable energy, AI enabled smart wind turbines could increase their output by as much as 8% by making intelligent adjustments to the direction of wind speed.
In the healthcare industry, device makers have already created AI-enabled implants for epilepsy patients that have reduced the frequency of their seizures. The development of similar kind of devices for other medical conditions could reduce emergency room visits and drastically lower health care costs. In agriculture, AI-enabled machines use their intelligence to determine where exactly to spray herbicide or fertilizer. The implementation of such machines could reduce the cost of farmers by 90%, but might be ominous for chemical companies as their demand might reduce.
Smart devices have the potential to help industries achieve new levels of efficiency and effectiveness, along with reducing material waste and cost of production. As machines learn from experiences, adapt to changing situations, and predict outcomes – the impact of pervasive AI is set to beyond being just faster, better, and cheaper.