Electronic Products & Technology

The future of AI – Three predictions for 2022

By Nick Romano, CEO – Deeplite   

Automation / Robotics Electronics AI artificial Editor Pick intelligence

Investments in AI will continue to increase over the next three years

Artificial intelligence, or AI, continues to play a greater role in improving product performance, how companies operate, and our everyday lives. According to McKinsey’s recent report The State of AI in 2021, nearly two-thirds of respondents said that their companies’ investments in AI will continue to increase over the next three years. In addition, the results show that it helps the bottom line too – 27 percent of respondents report at least five percent of earnings before interest and taxes (EBIT) that’s attributed to AI.

As companies continue to try to expand AI into new industries, products and applications, it will become more challenging to bring it to these emerging areas – especially on devices that lack the computing and power resources needed to effectively run it. In particular, small devices that are located far away from corporate data centers and operate on small batteries will need a new breed of AI that is smaller, faster and “lighter” than traditional approaches. More research will be needed, and new tools will need to be developed and improved to continue AI’s evolution and expansion. However, there are new advancements and innovations ahead that could create significant opportunities for the market.

Source: Getty Images

What are these advancements and opportunities? I’ve selected four below that could have a significant impact this year, and beyond. Here are my top four AI predictions for 2022.

  • The computer vision market will take off, driven by edge AI – There will be an explosion of market opportunities for computer vision applications as edge AI technology rapidly evolves and advances. Companies that move fast will take advantage of this technological evolution, as more (existing) devices will need to be upgraded with new AI (computer vision) functionality to remain viable and competitive. The fastest and most cost-effective solutions will use existing Arm processor hardware instead of deployments based on expensive and power-hungry GPUs (graphics processing units).

Many companies have enhanced their computer vision to do some interesting things – for example, factories are using it to automatically find defects in products on the assembly line. Another company has developed dashboard cameras for trucks that use computer vision (with AI) to read speed limit signs and dynamically slow the speed of the vehicle to avoid tickets and prevent accidents. While these are exciting technology advancements that can help improve efficiency and safety, there are many more opportunities ahead for computer vision in markets like healthcare, automotive, consumer electronics, security and surveillance, and more.

  • PyTorch will emerge as a leader – While the debate on PyTorch vs. TensorFlow continues, PyTorch will rise to become the leading platform for AI model innovation in the research community. It will become the best way to access advanced performance when building edge AI models for computer vision. Several top companies are already using PyTorch, including Amazon, AMD, Intel, NVIDIA, Qualcomm and others, but many more will join this list in the future.
  • TinyML will be widely embraced – As AI continues to impact nearly every facet of our lives, organizations will be looking at ways to incorporate this technology into ultra-low-power devices at the edge. TinyML, an eco-friendly, fast-growing field of machine learning, will be embraced by organizations looking to run AI on battery-operated devices, while also keeping sustainability in mind. Think of Google Home and Amazon Alexa – tinyML is required to get the AI on those home devices. The tinyML Foundation is working with several AI sponsors to spread awareness and education about this field and get the younger generation (as young as high schoolers) of emerging AI developers involved.
  • Deep learning/AI talent heads for the startups – As part of the “great resignation,” AI talent, including those with coveted deep learning experience, will leave their high paying jobs at large, big-brand-name companies and migrate to startups, looking for the opportunity to work on cutting-edge projects, learn new skills, and work with innovative technologies. This type of mobility may be accelerated by new working dynamics in the COVID-19 area, including remote work, which can make it faster and easier for people to change companies regardless of their location. According to a Northeastern University survey of more than 1,000 C-level executives, 35 percent of organizations are leveraging remote work to expand their talent pool.

In all, when looking at the next frontier for AI, the edge presents significant opportunities but will need some advancements to help bring computer vision to more applications in more industries. Innovation will not only expand the reach of computer vision, but will help companies to do more with their existing equipment, protecting their investments and avoiding expensive equipment replacement costs. More powerful, but “smaller” AI can bring benefits to people, companies and entire industries in 2022 and beyond.


Nick Romano, CEO – Deeplite


Stories continue below

Print this page

Related Stories