Abstract

The dependence of structural physical systems including Supervisory Control and Data Acquisition (SCADA) systems on AI technology is growing rapidly. The mandate of AI to achieve efficient and effective industrial supervisory systems is clear; however, threats resulting from both internal malfunctions and external cyber sabotage have become of crucial concern to SCADA systems that depend on AI. AI defence mechanisms are often installed within system architecture and through other external sources. Careful balancing of defence mechanisms to counter attacks and overcome systems vulnerabilities remain critical to the very protection of public lives which have become progressively reliant on these physical systems. In these developments, the passive role of human governance of technology systems cannot be wholly exempted. This paper employe the action research strategy with primary focus on two SCADA control rooms in the Emirate of Abu Dhabi. These control rooms handle over 60% of all non-law enforcement physical and infrastructural systems monitoring across the Emirate of Abu Dhabi. Operational AI integrated SCADA Systems are diagnosed with the help of document analysis and informal interviews. Action planning entails careful matrix modelling of possible courses of actions that balances AI defence and attacks in the SCADA environment. The implementation, evaluation, and specification of learnings are reserved for the second part of the present paper. The results of the first two stages of the action research are critically discussed to reveal evidence on how AI is operationalized in SCADA monitoring, data collection, and control centring.