- Experts consensus says that artificial intelligence is the next technology that will sweep through every industry, just like cloud did, from high tech to retail to agriculture to supply chain. The problem with AI implementation, however, is that the technology is not evenly or widely distributed within or across industries, said Dr. Keith Dreyer, vice chairman of radiology at Massachusetts General Hospital and associate professor of radiology at Harvard Medical School, speaking Thursday at the GPU Technology Conference in D.C.
- Futuristic and niche use cases of AI, machine learning and deep learning are not hard to find. Dreyer showed how combining AI with digital imaging can improve diagnosis precision in healthcare. Image classification — such as facial recognition and numbered identification of humans and cars — in policy body cameras and vehicle-mounted cameras demonstrated law enforcement applications of the technology. Some are even using AI in farm equipment to automate cotton weeding.
- While more than 40% of companies have AI experiments or pilot projects, only one in five have deployed AI at scale or for a core business function, according to a Harvard Business Review survey of 3,000 executives.
While cloud already serves as the technology backbone for many companies, AI will be more difficult for companies to include in digital transformation efforts.
That being said, digitization in general is still behind the curve, and estimates have the average business across sectors is about 40% digitization, according to McKinsey. The high tech industry has reportedly passed the halfway benchmark in digitization efforts.
So how is the business world, much of which is still catching up on digitization and the cloud, supposed to bring advanced AI applications into their business models — let alone when some have only just acquired wireless internet?
Efforts at democratization have received significant focus in recent months, and tools such as drag-and-drop AI algorithms, accessible AI platforms, training programs, automation of basic AI work and other vendor offerings are helping to bridge the gap between AI in theory and practice.
AI talent finds itself in a drought even more severe than most technical skills. The shortage is rocketing up salaries to the six-figure range and pulling AI experts out of academia — inciting worries on education institutions' ability to train the next generation of AI talent. Educators still in place have expressed concern that the tech market is pushing AI focus into "curiosity-driven" advanced and niche applications rather than basic research.
For some industries, an opaque understanding of how AI and ML will mesh with regulations and laws offers another challenge. Fintech in the financial services industries received a warning about the rise of third-party dependencies and new players outside the scope of existing regulations from the Financial Stability Board on Wednesday.
While the heights of AI and ML applications are certainly worth the wonder they incur, early and easy to implement applications of the technology cannot be left behind. For tech giants with the ability to spend billions or trillions on these sectors, amounts that dwarf spending in virtually every other industry, much of this onus remains on them.