This week’s Industrial Strategy policy paper from the UK government places Artificial Intelligence right at the top of the list of ‘Grand Challenges’:
“Artificial intelligence and machine learning are general purpose technologies already starting to transform the global economy,” it states. “They can be seen as new industries in their own right, but they are also transforming business models across many sectors as they deploy vast datasets to identify better ways of doing complex tasks – from helping doctors diagnose medical conditions more effectively to allowing people to communicate across the globe using instantaneous speech recognition and translation software.”
I find the prospects for AI and machine learning tremendously exciting, but I want to try to draw a distinction between the strategy’s characterisation of these technologies as better ways of doing complex tasks, and the Augmented Intelligence approach which enables humans to make optimal decisions in the most complex environments.
Artificial intelligence is well suited to reviewing huge historical data sets and making predictions based on the correlations in that data. Machine learning models can help users understand the states through which an asset transitions on its road to failure, and in doing so make predictions about the probability of failure at some future point. Those predictions can be hugely valuable, but as a human being who must make a decision about that asset, I need to place that prediction in the context of multiple dimensions such as finance, criticality, human resources, regulations and laws, and other organisational objectives and constraints.
In short, the decisions I make are not made in a vacuum but in a real and complex world. Making the optimal decision therefore requires an understanding of how that decision will impact, and be impacted by, interconnected and interdependent systems within and external to my organisation.
Making optimal decisions in this context means seeking Augmented Intelligence, not artificial intelligence. Augmented Intelligence is based on combining multiple and various expert resources with complex systems modelling to power simulations that can predict possible emergent phenomena – new behaviours or scenarios that arise from the interactions of different systems – as well as the cascading effects of each and every decision throughout the entire interconnected, complex system.
Augmented Intelligence and Resilience
In Systems thinking: an approach for building resilience in PR19, I wrote about Ofwat’s Resilience in the Round paper and how this advocates a systems thinking approach to build resilience in water and wastewater utilities. The UK Government’s guide to improving the resilience of critical infrastructure identifies four aspects of resilience: redundancy, resistance, reliability and response/recovery.
It’s this final aspect of response and recovery where Cosmo Tech has recently helped a client apply an Augmented Intelligence solution to improve resiliency in a crisis.
The Syndicat Des Eaux D’ile de France (SEDIF), a water network owner in the Paris region, wanted to understand how they could prepare themselves for a major emergency on their network, such as the accidental or intentional contamination or breakdown of a major production site, and how they could manage such an event in real time. By modelling the water system of the French capital, identifying critical consumption points, distribution points, and pumping stations, as well as the human resources required to service the water network in times of emergency, and, crucially the inter-dependencies between these systems, planners can run simulations in order to determine the optimal response to any water infrastructure emergency, either before the event or in real-time when under real constraints.
Armed with Augmented Intelligence, SEDIF is able to respond optimally to an emergency not based on correlations extracted from huge historical data sets but rather from accurate models of their real systems. There’s a video of the solution in action here, along with some more information about how Augmented Intelligence can help water and wastewater utilities make optimal decisions as they work to improve resilience in their businesses.