Automation, artificial intelligence and robotics: emerging issues and debates in human factors and ergonomics

1. The Ironies of Automation

Forty years ago, Lisanne Bainbridge published a brief paper in the journal Automatica, titled Ironies of Automation (Bainbridge 1983). It is likely that Bainbridge, at the time a cognitive psychologist at University College London, did not anticipate that the ideas she presented would become so influential. This special issue collection was prompted by an opportunity to recognise the 40-year anniversary of this important work. The paper argued that automation, instead of solving problems, can in fact introduce new issues, as operators lose skills and the capacity to intervene when automation fails. It is often referred to within the field of Human Factors and Ergonomics (HFE) in response to calls for ‘fully automated’ systems without human involvement, including artificial intelligence (AI) based systems, and their embodiment within robotic technologies.

Various authors have revisited the Ironies of Automation over the years. At 30 years post-publication, Gordan Baxter and colleagues (2012) concluded that the ironies remained relevant in the context of modern, complex technological systems. In particular, the authors reflected on two issues; one being skill degradation and the other, the difficulties associated with human monitoring of sophisticated technological systems in real time. A few years later, Barry Strauch (2018) wrote a tribute paper aligning the Bainbridge’s ironies with automation-related accidents that have since transpired, as well as identifying new ironies that have since emerged.

Citation counts can also provide evidence of influence, at least in academic circles. 40 years post-publication, the Ironies of Automation is generally considered to be one of the most highly cited papers within HFE. At the time of writing, the paper has been cited 1185 times according to Scopus, and 2826 times according to Google Scholar. Based on Scopus data, Figure 1 shows the trend of citations over the last 15 years, suggesting a period of growth between 2009 and 2019. And while the growth may be slowing, citation numbers remain relatively high at approximately 70–80 citations per year over the last 3 years (noting that 2023 figures are incomplete).

Figure 1. Scopus citations for the Ironies of Automation (Bainbridge 1983).

2. Current trends in technological advancement

Technology is continuously advancing, yet the last 12 months has seen some noteworthy breakthroughs. Specifically, these include significant new advancements in AI. The release of chatbots based on large language models (e.g. Open AI’s ChatGPT) and AI-based image generators (e.g. Midjourney, which made headlines after being used to create the winning entry of Colorado State Fair’s annual art competition), have increased public attention and interest in the topic. AI and automation are no longer the purview of industries such as advanced manufacturing or autonomous shipping; instead, direct use is becoming the norm for workers, students and the general public. Until recently, members of the San Francisco community were able to hail automated ‘robotaxis’, operated by General Motors Cruise, on public roads with no safety driver on-board. However, a series of safety incidents led to the suspension of testing permits. There has also been increased interest and work towards new regulations for AI. For example, at the time of writing, member states of the European Union (EU) are negotiating the details of the AI Act, which aims to ensure that AI systems are safe, transparent, traceable, non-discriminatory, and environmentally friendly. In addition, the Bletchley declaration was recently signed by 28 countries and the EU, affirming that ‘for the good of all, AI should be designed, developed, deployed, and used, in a manner that is safe, in such a way as to be human-centric, trustworthy and responsible’ (Department for Science, Innovation and Technology 2023). Clearly, HFE has a vital role to play in supporting the achievement of these aims.

3. Purpose of this special issue collection

With the intention of acknowledging 40 years since the publication of the Ironies of Automation, and in light of continuing technological acceleration, the aim of this special issue was to provide a forum for highlighting current knowledge gaps, debates, and controversies relating to automation, AI and robotics and to inspire future developments in HFE to address the emerging challenges. The resulting collection of papers is diverse and provides a range of contributions and insights. Some authors directly revisit the ironies of automation in the age of AI (e.g. Endsley), others reflect on specific domains such as healthcare (Zayas-Caban et al.), defence (Schraagen), and road transport (e.g. de Winter et al.). We have papers discussing theory (e.g. Naikar et al.) and applying HFE methods (e.g. McLean et al.), while others take a step back and question the implications for long-standing assumptions of the HFE discipline (e.g. Hancock).

Although we did not formally invite ChatGPT to contribute a paper, it seemed fitting to ask the generative AI model for its thoughts. We prompted ChatGPT 3.5 with the topic of the special issue. The response is provided in Table 1. The result, while somewhat generic, provides a reasonable summary of several key issues, and promisingly, provides caveats and advice to the reader to check the content, including via reference to publications in the field.

Table 1. ChatGPT 3.5 response to the special issue topic.

Prompt

Response

Within the human factors and ergonomics field what are the current knowledge gaps, debates, and controversies relating to automation, artificial intelligence and robotics?

As of my last knowledge update in January 2022, I can provide insights into some general trends and discussions within the human factors and ergonomics field regarding automation, artificial intelligence (AI), and robotics. However, it’s important to note that the field is rapidly evolving, and new developments may have occurred since then. Here are some key points: 1. Trust and reliance on automation: Knowledge Gap: Understanding how users develop trust in automated systems and how this trust influences their reliance on automation is an ongoing area of research. The dynamics of trust are complex and depend on factors such as system reliability, transparency, and user experience. Debate: There is a debate about the appropriate level of automation in various domains. Striking the right balance between human control and automation to optimise performance and safety is a challenging aspect. 2. Human-AI collaboration: Knowledge Gap: Research is exploring how humans can effectively collaborate with AI systems. This involves understanding the optimal division of tasks between humans and AI and designing interfaces that facilitate seamless collaboration. Debate: There are discussions about the potential impact of AI on job roles and whether automation leads to job displacement or the creation of new, complementary roles. 3. User interface design: Knowledge Gap: Designing user interfaces for complex systems involving AI and robotics requires an understanding of how users perceive and interact with the technology. This includes issues related to information presentation, decision-making support, and the integration of AI suggestions. Controversy: Balancing the need for comprehensive information with the risk of information overload is a point of contention. Designing interfaces that provide relevant information without overwhelming the user remains a challenge. 4. Ethical considerations: Knowledge Gap: Ethical issues related to AI and robotics in human factors include issues such as bias in algorithms, privacy concerns, and the ethical implications of decisions made by autonomous systems. Controversy: There are ongoing debates about the ethical responsibilities of designers, engineers, and organisations in ensuring that automated systems are fair, transparent, and unbiased. 5. Training and skill development: Knowledge Gap: Understanding how to train individuals to work effectively with automated systems is a key area of research. This involves assessing the impact of training programs on user performance, situation awareness, and overall system understanding. Debate: There are discussions about the types of skills that will be most valuable in a future where automation and AI are prevalent. This includes debates about the role of education and training in preparing individuals for this changing landscape. 6. Psychological and behavioural aspects: Knowledge Gap: Exploring the psychological and behavioural aspects of human-robot interaction, including user preferences, attitudes, and the emotional impact of interacting with AI and robotic systems. Debate: Discussions focus on how emotional intelligence in machines can enhance human-machine collaboration and the potential implications for user well-being. Research and discussions in these areas continue to evolve as technology advances and the deployment of AI and robotics becomes more widespread. To get the latest information, it’s advisable to check recent publications and conferences in the field of human factors and ergonomics.

4. Summary

HFE is uniquely placed to understand the potential effects of automation, AI, and robotics, both in the workplace and in society more generally. Our strong history in analysing and designing human-technology interactions and understanding wider sociotechnical systems impacts should put us at the forefront of discussions of new and emerging technologies. Therefore, it is increasingly important for us to apply (and extend) our theories and our methods to support the design of advanced technologies and the larger systems within which they will be integrated. We hope that the papers in this special issue collection will inspire research and practice that supports the design of safe, efficient, effective, and ethical automation, AI and robotics.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.