There are talks about advanced robotics in manufacturing/warehouse, driverless cars, IoT. But how meaningful are the changes about the automation of knowledge work? The automation of knowledge work is #2 in the disruptive technologies only after mobile internet.
The knowledge workers are various professionals in technology, healthcare, legal, finance and science sectors.
In the information technology they could be web designers, researchers, programmers, system analysts, technical writers. In the other fields, knowledge workers are physicians, pharmacists, scientists, engineers, building architects. You could also include public accountants, attorneys, and thinkers in the social and economic sectors. To increase efficiency of their work, the specialists would need to acquire help from AI tools and map their processes accordingly.
Mapping work processes of SME – Subject Matter Experts – by the AI tools drive data more efficiently than any prior tech methods. They mimic the work processes of human experts. The types of experts are relatively broad and contain scarce data in their category, they don’t generate huge data quantities.
When they started, the machine learning and deep learning with AI apps have been founded on massive data to train and build systems. The approach has been the bottom up. Since the focus shifts to invaluable data however low volume, AI systems will engage top-down approaches more often. The emphasis while acquiring data and training, will give the AI the capacity to capture, incorporate and use the knowledge workers’ specialized expertise.
Within the past decade, AI-driven systems that detect images, NLP (Natural Language processing), and voice recognition have developed with higher accuracy. Data becomes more important for the variety of AI tools created on numerous types of algorithms. They have become more efficient lowering – to below 3 percent – the error to recognize images correctly.
In other words, the accuracy of detection of thousands of images by AI has raised to 97.3% which represents an increase of over 35 percent. The AI algorithms outperform current knowledge workers’ capacity to detect images, sounds and other data employing higher efficiency. At the Laboratory of Medical Image Processing, Brest Hospital University, France the AI-driven tool has correctly recognized the medical devices utilized by a surgeon in a minimally invasive cataract surgery. The AI machine winner in the vision system oversaw 50 videos of cataract surgery detecting correctly the tools that enable medical personnel to use during operations.
Improvement of AI recognition systems have potential applications in many areas that support the activities in the operating rooms: surgical training, report generation, and even suggestions to decision-makers. They also have applications in developing technology, science, training, and all sectors where knowledge workers undertake responsibility.