Responsibility Without Accountability: The Paradox of Responsible AI in the BPM

Overview

The discourse of "Responsible AI" has proliferated rapidly across industry, regulation, and academia, with principles such as fairness, transparency, accountability, and human oversight now embedded in frameworks like the EU AI Act, NIST AI RMF, and Australia's AI Safety Institute guidelines. However, a growing body of critical research identifies a significant gap between these stated principles and actual organisational practice, as well as what has been termed "ethics washing" or "AI governance theatre". In the BPM domain specifically, this paradox is acute: organisations are incentivised to automate processes for efficiency and cost reduction, yet the Responsible AI obligations of explainability, human oversight, and impact assessment directly increase process complexity, costs, and friction. The transition from AI governance as a set of principles to enforceable operational practice is now a core enterprise risk function with legal, financial, and reputational consequences; yet BPM scholars have not adequately theorised how Responsible AI principles can be operationalised within the constraints of real-world process management. The negative consequences of this failure include unmonitored automated decisions affecting citizens' rights, workforce displacement without adequate process safeguards, and AI systems whose adverse impacts are discoverable only after harm has occurred.
RQ: Why do organisational BPM practices fail to operationalise Responsible AI principles despite policy commitments, and what institutional, technical, and cultural mechanisms must be activated to embed responsibility into the AI-augmented BPM lifecycle?

Research engagement

Systematic Literature Review, Qualitative Survey

Research activities

Systematic literature review (Preferred Reporting Items for Systematic Reviews (PRISMA)) synthesising research at the intersection of Responsible AI and BPM, mapping existing frameworks against empirical evidence of practice

Survey research across BPM practitioners and AI governance officers to empirically test the principle–practice gap across industries and organisational sizes

Critical Realist retroduction to identify the generative mechanisms (institutional pressures, regulatory ambiguity, skills deficits, incentive misalignment) that produce irresponsible process automation despite policy commitments

Benchmarking against the ISO/IEC 42001 alignment model and its three axes: governance and transparency, data and semantic infrastructure, and sociotechnical adoption factors

Research skills

This topic requires advanced skills in systematic and integrative literature review, including the ability to synthesise across multiple disciplines (AI ethics, BPM, organisational theory, and public policy). The student must be capable of designing and executing qualitative interviews and survey research, including instrument development, sampling strategy, and qualitative data analysis (e.g., NVivo) to empirically test the principle–practice gap. Grounding the study in Critical Realist ontology requires the ability to theorise generative mechanisms — such as institutional isomorphism, regulatory ambiguity, and incentive misalignment etc., that explain surface-level behaviours. Familiarity with maturity model development methodology (e.g., based on CMMI or BPM Maturity Model approaches) is advantageous for producing the Responsible BPM Maturity Model deliverable.

Outcomes

1. A Responsible BPM maturity model calibrating organisations' capacity to operationalise Responsible AI across process lifecycle stages
2. An empirically grounded CMO configuration (Critical Realist) explaining the principle–practice gap in Responsible AI–BPM integration
3. A policy-facing brief relevant to Australian AI governance (National AI Plan, AI Safety Institute) and international regulators
4. A journal or conference article targeted at AIS/BPM Conferences or a Q1 /Q2 Journal.

Skills and experience

Must have a good understanding of process modelling (IAB203) and information systems. diligence, creative thinking, and must be well motivated to enjoy the research topic,

Start date

2 November, 2026

End date

19 February, 2027

Location

QUT Y Block, Virtual

Additional information

Student researchers will receive hands-on training and support on research methodologies and qualitative data analysis.

Keywords

• Responsible AI
• BPM
• AI governance

Contact

• Dr Rehan Syed

0456036386

r.syed@qut.edu.au