Introduction
Agile has reached maturity in many organisations, and the success of Artificial Intelligence (AI) makes its use unavoidable for organisations to evaluate and enhance the performance of Agile teams. While AI promises a data-driven, objective lens for analysing team dynamics, productivity, and outcomes, its application to evaluate project team performance raises profound ethical questions—particularly when viewed through the lens of the Project Management Institute’s (PMI) Code of Ethics and Professional Conduct. This blog post explores the ethical implications of using AI to compare Agile team performance, guided by PMI’s values of responsibility, respect, fairness, and honesty.
Challenges
Responsibility: Ensuring Informed and Accountable Use
According to the PMI Code of Ethics, project managers are responsible for the decisions they make and their impact on stakeholders. The use of AI introduces a new layer of responsibility, as the algorithms and models deployed can influence critical business and technical decisions, team morale, and individual careers. Without a clear understanding of how AI systems operate, there is a risk of delegating accountability to technology, thereby abdicating human responsibility. Furthermore, Agile teams thrive on transparency and trust (Manifesto for Agile Software Development), and opaque AI models can undermine these foundational principles.
Respect: Preserving Dignity and Psychological Safety
The Agile Practice Guide and Manifesto for Enterprise Agility stress the importance of respect among team members. AI-driven comparisons risk reducing team members to data points, potentially diminishing their sense of value and psychological safety. This is especially concerning when performance metrics are derived from sources like code commits, velocity, or issue resolution rates, which may not capture the full scope of an individual’s or team’s contribution. Respecting people means recognizing their unique contexts, strengths, and the human factors that are not always quantifiable.
Fairness: Avoiding Bias and Ensuring Equity
PMI’s value of fairness calls for impartial and just decision-making. AI systems are only as unbiased as the data and assumptions upon which they are built. If training data reflects organizational biases or historic inequalities, AI may perpetuate or even exacerbate them. For example, if one team’s work is more client-facing and another’s more technical, or a team is developing new features while another is fixing defects, comparing them using the same metrics may yield unfair outcomes. The Manifesto for Agile Software Development emphasizes individuals and interactions over processes and tools, serving as a reminder that context matters—and fairness demands careful consideration of these differences.
Honesty: Transparency and Truthfulness in Reporting
Honesty, as defined by PMI, compels project managers to provide accurate and truthful information. The complexity of AI models can make it difficult to explain how conclusions are reached, leading to challenges in maintaining transparency. Agile values also highlight the importance of open communication and feedback loops. When AI-generated comparisons are presented without clear explanations, stakeholders may question their validity, eroding trust in both the technology and its advocates.
Recommendations
Human-Centric AI Design
AI tools should be designed to support, not replace, human judgment. Agile frameworks encourage continuous improvement, collaboration, and adaptation. In line with PMI’s ethical standards, organizations should ensure that AI complements human insight, enabling teams to learn and grow rather than simply ranking or scoring them.
Transparent Methodologies
Explainability is crucial. Both the PMI Code of Ethics and the Manifesto for Agile Software Development value transparency and trust. Organizations deploying AI to compare Agile teams must clearly communicate how data is collected, what metrics are used, and how results are interpreted. This fosters informed dialogue and minimizes misunderstandings.
Regular Bias Audits
Regularly audit AI systems for bias, in alignment with PMI’s fairness value. Engage diverse stakeholders—including team members—in the design and review of AI models. Use feedback to refine algorithms and ensure that they remain equitable and relevant.
Contextualized Metrics
Avoid one-size-fits-all comparisons. Agile teams operate in diverse environments, with varying goals, challenges, and customer needs. Tailor metrics to the context of each team, and supplement quantitative data with qualitative insights. This reflects the Manifesto for Agile Software Development's focus on individuals and interactions, ensuring a holistic view of performance.
Foster Psychological Safety
Use AI insights to support team development rather than to penalize. The Agile Practice Guide underscores the importance of psychological safety for learning and innovation. Frame AI-driven feedback as an opportunity for growth, not as a tool for surveillance or punitive action.
The Bottom Line
The integration of AI into Agile team performance comparison presents a unique set of ethical challenges. By grounding our approach in PMI’s values of responsibility, respect, fairness, and honesty—and by drawing on the wisdom of the Agile Practice Guide, Manifesto for Agile Software Development, Manifesto for Enterprise Agility, and PMBOK—we can harness the power of AI while upholding the highest ethical standards. Ultimately, the goal is not to let technology dictate our decisions, but to use it as a catalyst for deeper understanding, stronger teams, and more meaningful outcomes.
Question for Reflection: How can organizations ensure that AI-driven performance comparisons reinforce, rather than undermine, Agile and ethical values?
