The Ironies of AI: Why Smarter Systems Create Harder Human Problems
Human Factors has spent 70 years solving human-machine problems — here's why AI is both a new challenge and an old one in disguise.
There is nothing new under the sun… or is there? Recent years have seen staggering progress made in the field of Artificial Intelligence (AI), promising both prosperity and peril. This has left many organisations wrestling with the human impacts of this technology, with mixed success.
Enter Human Factors…
As a discipline we have been wrangling new technologies, and how people interact with them, for the past 70+ years. Building on decades of rigorous research on the interaction between humans and increasingly automated systems, Human Factors professionals are blazing a trail in the complex world of AI and autonomous systems.
Whilst AI creates some new Human Factors challenges, it also exacerbates some good old-fashioned ones. To deal with both the new and the old we must stand on the shoulders of giants, learning from the past whilst adapting our toolkits of approaches to novel problems… and perhaps adding a few new spanners where necessary!
A fantastic example of this is Mica Endsley’s publication “Ironies of Artificial Intelligence” published a few years ago. Yes, I realise that feels like ancient history given the lightning-fast pace of AI developments!
Building on seminal work in the field of human-automation interaction, Dr Endsley proposes five ‘ironies’ of AI:
Irony 1: AI is still not that intelligent.
Irony 2: the more intelligent and adaptive the AI, the less able people are to understand the system.
Irony 3: the more capable the AI, the poorer people’s self-adaptive behaviours for compensating for shortcomings
Irony 4: the more intelligent the AI, the more obscure it is, and the less able people are to determine its limitations and biases and when to use the AI.
Irony 5: the more natural the AI communications, the less able people are to understand the trustworthiness of the AI.
Are these ‘ironies’ useful? Do they capture the main human-related challenges of AI? Will they stand the test of time given new advances in Large Language Models, Generative AI, and Agentic AI? These are all questions Human Factors professionals are wrestling with as we face the age of AI.
UCD have extensive knowledge and experience helping organisations grapple with the human impact of AI and autonomous systems. If you are interested in finding out more about UCD’s capabilities in Human Factors and AI please don’t hesitate to get in touch.