Siri Padmanabhan Poti

Research Student - PhD

Research Program: Biomedical & Human Technologies
Research Lab: Human Machine Interaction [[link]]

Thesis Title

Enabling Effective, Ethical and Meaningful Human Collaboration with Autonomous AI Systems

Research Project

In Industry and society, humans are increasingly collaborating with autonomous artificial intelligence (AAI) systems. This research adopts a human-centric, socio-technical philosophy to examine how successful human collaboration with safety / mission-critical AAI systems may be enabled. Success is predicated on the effectiveness, ethicality and meaningfulness of the collaboration, thereby enabling legal and ethical accountability, performance effectiveness and risk reduction. The research employs five methods: i) scoping review of 163 primary peer reviewed sources (2020–July 2024), to synthesise the ontologies of elements and constructs related to the collaboration, ii) theoretical integrative review to model the key augmenters, detractors and enablers in the meaningful and effective human collaboration with AAI systems iii) conceptual modelling of transformation in the system development lifecycle to enable explainability, interpretability and assurance from AAI systems, iv) conceptual modelling of explainability and trust through a cyber-physical intermediary, and v) qualitative evidence synthesis to broaden the understanding of gaps in ethics mechanisms for AAI systems. This research makes novel conceptual and theoretical contributions  by conceptualising a ‘Collaboration-by-design’ framework with six ‘enablers’: i) methodological interventions for transformation in AAI system development lifecycle to enable affordances for governance ii) conceptual model of Mediating explainer (MeX), a cyber-physical intermediary for explainability and trust in collaboration, iii) anticipatory general ethics library (AnGEL) for enabling trustworthiness  iv) adapted Cognitive Reflection Test dual-task-interaction-testing protocol, with prospective empirical components, v) thought training and vi) design matrix as a boundary object. These have implications for policymakers, cross functional practitioners, multidisciplinary teams, and organisations in emerging technologies.

Qualifications

• Master of Business Administration, Griffith University (2017)
• Master of Research (School of Computer, Data and Mathematical Sciences, Western Sydney University
  (2020)

Contact