Part 1: Introduction Chapter 1: Introduction to the Future of Work and Automation The introduction sets out the context and contents of the book. It introduces automation and robotics as a technology and AI as a core technology in those areas. The authors are engaged in several research projects that will tackle some of these fundamental challenges. The research is then placed into the context of the future of work. The possible outcomes of each research project are placed in a spectrum of automation for the future. The changing nature of work in the future is outlined as a potential outcome of the changes that are thought inevitable with the increasing automation of existing tasks and jobs. The scene is set for the subsequent introductory chapter on the current thinking about the automation/human relationship and the changes that are likely to happen. Chapter 2: Are Robots Replacing You? There are many conflicting arguments about the impact of automation on the working population and even on the impact on government and individual''s economies.
This chapter discusses some of the theories and predictions from commentators and analysts who have interesting and often different views of the next 10 or 20 years. There is no doubt that automation will be disruptive, displacing people from many tasks, but there is some doubt that it will be destructive, replacing people from many jobs and leaving a world with massive unemployment. The next section discusses the technology challenges that this changing world is influencing. Part 2: Technology Challenges Chapter 3: Rules that Don''t Work, Bots, and Chatbots Although applications have long been available that endeavour to mimic a human response to enquiries or questions these applications take a considerable time to develop because of the need to map all of the conceivable responses to any question that can be posed. A recent advance in the automation of tasks is the development of the AI software robots or bots. These software robots, used most frequently in a support service capacity, mimic the style of an online or telephone conversation that would be expected from a person. There are complex AI decisions that drive the interaction between bot and human, but the most crucial development is the bot''s ability to learn and improve responses to their interlocutor. Chapter 4: Robotic Process Automation Many processes can be automated and have been in the past.
Early IT based automation were in the area of automating simple ledger systems and process, initially in financial systems. Robotic Process Automation (RPA) is the most current methodology that can automate complex business processes. The combination of process analysis, process mining and AI tools that can learn is making an impact. The ability to develop fast and repeatable processes that deliver decisions is transforming the financial world. Insurance and Banking organisations are using RPA to personalise their products and it is becoming more difficult to tell if the voice, text or decision you get from an insurance application is from a human or a machine. Managing the risk of a poor or opaque decision-making process will be a feature of this type of automation in the future. The ALOHA project is investigating this aspect of machine learning and AI management. The need to integrate complex data and ensure that the data is not biased is demonstrated and the challenges will be examined in more detail in Chapter 7; Data Fusion.
Chapter 5: Robots, Collaboration, and Collaborative Robots The world of Robotics is very broad and takes in industrial robots, human augmentation, and collaborative robots. This chapter will have a focus on the challenges of Collaborative Robots (COBOTS) since they are the subject of a research project sponsored by the Centre for Visual and Decision Informatics. COBOTS can range from manipulators that are moved by humans to robots that carry out generalised tasks in collaboration with one or many humans or robots. In this chapter we outline the challenges in collaborative robotics and we note the technology that is being researched to resolve those challenges. Particular emphasis will be placed on; data fusion, incomplete or missing data analysis, real-world views, policies for safety, collaboration and continuous risk management. Chapter 6: Smart Buildings and Special Cases A challenge for an automated world and the future of work is how the built environment will respond to new tasks, jobs, employees and their automation. Intelligent buildings frequently include older sensors and actuators that need to be reconciled with new equipment that is retro fitted. The data fusion needed to develop an accurate picture of the whole building is added to the increasing personalisation that smart buildings promise.
There are also special cases, such as Smart Cities and Autonomous vehicles that demonstrate many of the same characteristics as an intelligent or smart building. The data and policy challenges of autonomous vehicles are often similar to those of COBOTICS and will be discussed here. The Tieto Empathic Building plans are outlined with their accompanying requirements. Part 3: Technology Research Chapter 7: Data Fusion One of the fundamental building blocks of automation is the expansion from digital data analysed by AI into a far more complete "real world" view that includes associated audio and video data. This is of great interest in robotic implementations in general and it is a vital component of collaborative robotics and smart buildings. The decisions based on data to be analysed requires fast, near real-time responses and the decisions frequently need to be made by low power devices at the edge of a network. The EU funded project "ALOHA" will research the implementation of support for diffused low energy and heterogenous architectures. The focus of the project will be on the effect of Multi-Processor System on Chip, dedicated accelerators for deep learning tasks and Field Programmable Gateway Array hardware.
The potential to deliver low power machine learning on these style of components at the edge of the network can have an important effect on personalisation and management of intelligent buildings and COBOTS. Chapter 8: Real-World Common Models Each member of a collaborative task uses a difference mental model. Humans, drones and robots all have their individual model of the reality that they see and there must be a common agreed model that enables them to interact. This chapter discusses the common areas of individual real-world models and how collaboration is enabled using data fusion as a core technology. Data fusion is also discussed in the context of data poisoning and model bias using examples from the ALOHA project. Chapter 9: Collaboration and Policy A feature of increasing automation will be the policies and tools for collaboration between automated processes and robots. Policies will need to be flexible, easily implemented and updated by changes in the risk profile of the processes and robots. In this active world safety will be of primary importance and risk assessments will need to be real-time and continuous.
Current risk management and decision making tends to be separate with Risk management being a one-off activity at the start of a project and decision-making being encoded in software using a set of rules that seldom change and fail to react to a changing risk environment. Policies for collaboration and in particular the complexities of machine to machine or machine to human task handovers. Part 4: Conclusion Chapter 10: Work in the Future This chapter reviews the technology challenges and research discussed in detail in previous Chapters. From this and the introductory section conclusions are drawn on the further challenges, the impact on automation and automation''s impact on work in the future. This chapter makes the point that tasks rather than jobs are most likely to be automated. Jobs will change to include a higher degree of automation and both jobs and tasks will result in displacement of workers rather than unemployment of workers. The workers themselves will have to learn new skills to manage automated tasks and react creatively to a new environment. Although Artificial General Intelligence is not discussed in the main body of the book it is mentioned here for completeness and its role in future automation is discussed.
The final question in the conclusion will look at the likelihood of automation and future work being influenced by artificial general intelligence in the medium term.