How to accelerate the uptake of AI in danish companies
Today, AI capabilities seem critical for companies’ ability to be competitive. Only few companies will be able or willing to implement large scale AI functions in the near future. So what can governments, investors, universities and others do to help accelerate the establishment of such functions and thereby the uptake of AI in companies?
Most Danish companies primarily work within the lower layers of the best practice model for AI functions. We believe the situation could be rectified by building shared services for the two or three top layers of the best practice model.
A. They could make Deep Learning Scientists available for companies in an easy and structured manner. Today, Deep Learning Scientists are largely non-exiting in Danish companies. They are primarily found in small groups at universities that are difficult to access for most companies.
B. They could facilitate external AI teams with Machine Learning Engineers becoming more easily available to companies, e.g. through subsidised trial projects. They could also help companies by training some of their existing employees, e.g. their best Data Engineers, to take on the Machine Learning Engineer role. Today, only few companies have Machine Learning Engineers on board and even those companies seldom know how to use them in a structured and focused way, but ask them to do tasks within other layers of the model. As a consequence, their skills quickly become obsolete, or they leave.
C. They could potentially also help companies build the capability to manage data needed for AI solutions. This involves collecting, storing, cleaning and calling data. It is often the most time-consuming activity and can take up 60-80% of the total time and man power. Today, even this initial step to get started with AI is a challenge for many companies.Today, Deep Learning Scientists are largely non-exiting in Danish companies.
We believe the companies that have access to shared services for the top layers in the model – we call it “Deep Learning as a Service” – will have a competitive advantage in the coming years.