Introduction
Agile organisations are at the forefront of innovation and adoption of new, modern tools and ways of working. Although Artificial intelligence started five decades ago, nowadays become the next big thing after Agile Transformations. As artificial intelligence (AI) becomes deeply embedded in organizational processes, its presence is not always visible to stakeholders. Hidden AI involvement refers to scenarios where AI influences outcomes, decisions, or operations without clear disclosure or transparency. While this technological integration can drive efficiency and innovation, it also raises profound ethical and governance challenges. This blog post examines these issues through the lens of the PMI Code of Ethics and Professional Conduct and foundational structures including the Agile Practice Guide, ISO 31000, and the PMBOK® Guide.
Challenges
1. Lack of Transparency and Accountability
The PMI Code of Ethics emphasizes honesty, responsibility, and respect as cornerstones for professional behaviour. When AI systems operate in the background without explicit acknowledgement, it undermines these values. Stakeholders may be unaware of AI’s role in decision-making, leading to a lack of accountability for outcomes. This opacity can erode trust and hinder effective governance, as highlighted in ISO 31000’s call for risk communication and stakeholder engagement.
2. Difficulty in Tracing Decision Logic
A recurring challenge discussed on the Agile forums is the “black box” nature of many AI systems. When the logic behind AI-driven decisions is concealed, organizations face difficulties in auditing, error correction, and compliance monitoring. Such opacity can exacerbate ethical dilemmas, especially when outcomes impact individuals or communities in significant ways.
3. Compromising Agile Values
The Agile Practice Guide prioritizes collaboration, transparency, and customer involvement. Hidden AI involvement can conflict with these principles by creating invisible actors in collaborative processes. This can undermine team cohesion, informed consent, and iterative feedback loops, all essential for adaptive governance.
4. Inconsistent Application of Professional Conduct
According to the PMI Code of Ethics, practitioners must “make decisions and take actions based on the best interests of society, public safety, and the environment.” If AI’s influence is hidden, professionals may inadvertently violate these principles, as their actions may be guided by unseen factors beyond their control or comprehension.
5. Risk Management Blind Spots
ISO 31000 and the PMBOK® Guide emphasize the importance of identifying and managing risks. Hidden AI introduces blind spots in risk assessment, as unidentified AI components can produce unforeseen vulnerabilities, propagate bias, or enable unintentional non-compliance with regulations and ethical norms.
Recommendations
1. Embrace Transparency as a Core Principle
Organizations should disclose the presence and scope of AI involvement in processes, especially when outcomes affect stakeholders. This aligns with the PMI Code of Ethics’ call for honesty and respect, and with ISO 31000’s guidance on communication with stakeholders.
2. Adopt Explainable AI (XAI) Practices
Leverage explainable AI frameworks to ensure that decision logic is traceable and understandable. This supports auditability, fosters trust, and meets the PMBOK® Guide’s requirements for transparency and documentation. Agile practitioners advocate for systems that allow users to question and understand AI behaviour, promoting ethical alignment.
3. Foster Ethical Awareness and Training
Regularly train teams on ethical considerations surrounding AI, referencing frameworks like PMI’s Code of Ethics, Agile values, and relevant governance standards. Ethics must be an ongoing conversation, not a one-time checklist.
4. Integrate AI Governance into Risk Management
Update risk management processes to recognize hidden AI as a unique risk category. The PMBOK® Guide and ISO 31000 advise proactive risk identification, assessment, and mitigation. This includes periodic reviews, stakeholder consultations, and scenario analyses focused on AI-driven processes.
5. Strengthen Agile Feedback Loops
Ensure that agile ceremonies and feedback mechanisms explicitly consider the role of AI. The Agile Practice Guide stresses the importance of transparency and continuous improvement. By surfacing AI involvement, teams can collaboratively address issues and uphold agile values.
6. Establish Clear Accountability Structures
Define roles and responsibilities for AI oversight. The PMI Code of Ethics and PMBOK® Guide recommend clear accountability for project outcomes. Assign individuals or committees to oversee ethical AI use, investigate concerns, and engage with affected stakeholders.
The Bottom Line
Hidden AI involvement poses significant ethical and governance challenges in organizations pursuing digital transformation. By drawing on established ethical codes, agile methodologies, and risk frameworks, leaders can foster transparency, accountability, and trust. Recognizing and addressing the risks of concealed AI is not only a compliance imperative but a professional and societal responsibility. The journey towards ethical AI governance is ongoing, demanding vigilance, education, and a commitment to core values.
Questions for Readers
- Has your organization encountered situations where AI’s influence was not fully disclosed? How was it handled from the ethical point of view?
- What ethical guardrails can teams use to ensure AI involvement is transparent and accountable?
- How might the ethical risks of hidden AI evolve as technology advances?
