This content was originally published on The Resilience Shift website. The Resilience Shift, a 5-year programme supported by Lloyd’s Register Foundation and hosted by Arup, transitioned at the end of 2021 to become Resilience Rising. You can read more about The Resilience Shift’s journey and the transition to Resilience Rising here.

Transformative technology is a topic we can’t ignore and, at the Resilience Shift, we want to put a resilience lens over its application to critical infrastructure. We attended an ENCORE Network workshop in June that focused on Artificial Intelligence (AI) for Infrastructure Monitoring.

In spite of the revolution that seems to be round the corner, we should be wary of the AI hype – something Gartner captures annually in their hype cycle of technology.

Gartner's Hype Cycle (Jeremy Kemp, Wikipedia, CC-BY-SA 3.0)
Gartner’s Hype Cycle (Jeremy Kemp, Wikipedia, CC-BY-SA 3.0)

AI is everywhere and at the top of the hype cycle for several technologies, such as machine learning and autonomous vehicles. This becomes clear in the latest available version of Gartner’s Hype Cycle (2017) and in their Top 10 strategic technology trends for 2018.

Gartner's Hype Cycle for Emerging Technologies 2017
Gartner’s Hype Cycle for Emerging Technologies 2017

One of the messages to take home from the ENCORE network workshop is that our society expects a computer to do better than humans, but it seems clear that there are many situations where this won’t be the case. A group of experts might do better in decision making. For example, machine learning can fail in a different way than a human would do when solving a particular problem.

We need to combine what computers are good at with what humans are good at. In 5 years from now, AI will probably be used for situation support rather than making autonomous decisions for operators.

Data availability and trust is similarly a key issue here, as the resilience of AI systems depends on using good data. In this sense, a potential risk could be a malicious injection of data in the system. Minimising the risk of cyber attacks is essential in order to develop trust in collected data that will be used for AI systems.

We are keen to hear from readers of this blog about how you can contribute to support us on this journey – if this is your case, please get in touch with us.