Towards the Future of Work: Managing the Risks of AI and Automation

Abstract

Many believe in a vision of the future where almost all work is automated. A first step already underway involves Robotic Process Automation (RPA) technology, which firms use to automate standardized computer work. The larger step that needs to be taken towards this vision lies in connecting RPA to AI, so that Machine Learning (ML) algorithms can be used to automate human “intelligence” and decision making in companies.

Management research surrounding the concept of Intelligent Automation (IA) is nascent and spans multiple domains. This thesis consolidatesthe fragmented research landscape through a Systematic Literature Review to address four research questions: 1) What use cases are IA fulfilling? 2) Which ML algorithms and technologies are employed? 3) What risks are associated with IA? and 4) What risk mitigation techniques are there? The findings paint a picture of what is needed to advance the value that IA delivers to firms and shore up professional practices.

Results show that the bulk (66%) of cases centered on document processing and chatbots. ML models, tended to be uninterpretable, posing transparency and risk challenges. The systematic coding of 77 key sources yielded 36 risks that fell into eight clusters that are explored in depth. Corresponding risk mitigation measures covered far less ground, leaving many risks unaddressed. The risk registry derived in this thesis offers a starting point for a structured approach to managing emergent risks necessary for IA to deliver on its promise to improve work.