The Fourth Industrial Revolution (4IR).
The fourth industrial revolution (4IR) is driven by technologies such as: artificial intelligence (AI), big data analytics, deep neural networking, cloud computing, 3D printing, blockchain, quantum computing, and advanced automation (Schwab, 2017). Notably, it is how these technologies interact and enhance each other that has driven the exponential changes of 4IR.
As with previous industrial revolutions, the skills most sought after by employers in 4IR are changing dramatically (Leopold, Ratcheva, & Zahidi, 2018), with a key industry landscape being dubbed Industry 4.0 (Drath & Horch, 2014). Along with adjustments for Industry 4.0, the ubiquity of new technologies in our daily lives, such as facial recognition, the Internet of Things (IoT), and mobile supercomputing, signals substantial changes in our social systems. Since the term Industry 4.0 was coined in 2014 (Drath & Horch) and the Fourth Industrial Revolution was declared in 2015 (Schwab) there has been extensive discussion on both of these topics with exact definitions shifting as one would expect of emerging fields. My research takes the view that Industry 4.0 is a subset of 4IR but recognises that some may dispute this view.
Examples from the four industrial revolutions. Exact dates of each revolution are disputed and more likely overlap than have distinct boundary markers.
When my research commenced, the primary context I was concerned with was the global shift into 4IR. There are numerous perspectives on the development of 4IR skills in higher education, most practically focussed such as IT skills. Even as these skills are pushed outside of computer science and IT departments into short programs and ‘hacky-hours’ designed to improve technology skills in non-tech disciplines, the focus is usually very practical – learn to code, manipulate statistics, and use other high-performance computing technologies. Whilst the rhetoric at universities is to break down interdisciplinary barriers and bring philosophers, social scientists, educators, and other non-tech researchers into the fold, most seem to be doing the same things, in the same ways, they have done for years – albeit often with some spruced up branding around the approach. Unsurprising then that we have seen limited success of true interdisciplinary approaches to ethical AI technologies.
From 18 months of reading and thinking on the problem, I have come to the opinion that the heart of the 4IR readiness is responsible and ethical use of data and new technologies within a context that considers societal common good and global sustainability. Understanding the impact of one’s research in relation to technologies such as AI will become increasingly necessary for doctoral graduates in a wide variety of fields beyond computer science and IT. It is from this core perspective that my work grows.
A 4IR skill that has attracted a lot of attention over the last couple of years is the ethics of AI with institutes popping up in countries and universities globally. Just a few to mention are the Oxford Institutes for Ethics in AI, The University of Melbourne Centre for AI and Digital Ethics, The MIT Schwarzman College of Computing, The AI Now institute at NYU.
For doctoral graduates, that skill might relate to how AI may be used now or in the future that would amplify the impact of their research on the broader environment. That broader environment could be defined in many ways such as an organisation or industry sector, a community or society, or a local or global natural environment impacted by Anthropocene changes. System impacts quite often cross many disciplinary boundaries, particularly when considering ethical and sustainability effects. Unfortunately, despite continuing calls for interdisciplinary skills between the humanities, social sciences, and applied sciences, the ethics of 4IR technologies tend to be taught in discrete pockets of higher education. In many instances of doctoral education, the systemic ethical impacts of data framing and 4IR enhancements to research many not be explicitly addressed at all.
Research impacts from social science, law, natural science, psychology, medicine, economics, philosophy, and business – to name just a few – should in many cases consider their relationship to 4IR and AI technologies. These technologies are becoming inescapable and when data is framed and collected without consideration of the use of 4IR technologies now, once research exists it may very well be used in the future by well-meaning researchers that do use AI driven technologies. If we can consider now how our work may be used later, we have a better opportunity to build in fairness and sustainability into the original work. This last goes to the core characteristic of 4IR, rapid change and an ever-increasing need for agility and adaptation. Future proofing one’s work for 4IR requires development of a set of qualities beyond just practical IT skills.
Defining what are the most critical 4IR skills for doctoral students in the 2020s is an important starting question to answer in this thesis. Once that is answered I can ask, to what extent can doctoral researchers benefit from a better understanding of ethical AI and how it may intersect with their work?