DS Cohen & Associates
Artificial Intelligence (AI) has entered the HR suite, offering both promise and peril as the newest shiny object. It boasts speed, efficiency, and claims of fairness. Proponents celebrate the rise of evidence-based, bias-free hiring and talent management with objective metrics. Critics, however, fear a “black box,” an inscrutable system that may perpetuate old biases under the guise of progress.
Succession planning is where this paradox reaches its peak. Organizations attempt to forecast future leaders by scrutinizing present performance. With AI’s predictive analytics, the hope is to overcome hunches and favouritism with data-driven insights. Yet, the ethical and practical pitfalls can be greater than those in manual succession planning. The very automation that promises impartiality can also amplify historical prejudices and erode trust. As with any house built on shaky foundations, results will ultimately be unreliable.
The Shadow of the Black Box
Central to concerns about AI in succession planning is opacity. Ideally, algorithms would clarify what human managers have long obscured. In reality, most AI systems are black boxes. They are complex models whose workings even their creators cannot fully explain.
This isn’t hypothetical. For example, Amazon’s AI recruiting tool, trained on resumes from a predominantly male workforce, perpetuated gender bias. The lesson: AI trained on skewed data will reinforce existing prejudice. Organizations may thus persist in biased decision-making while believing themselves more progressive.
Despite this risk, organizations are drawn to AI’s data: performance reviews, engagement metrics, leadership program completions, and learning agility scores. These sources identify “high-potential” or “ready-now” leaders. But are these recommendations’ objective, or do the algorithms merely reflect hidden preferences embedded in the training data?
Data Quality: The Crux
The value and danger of AI lies in the quality of its data. Performance reviews, a common input, are notoriously unreliable. Gartner found that 79% of HR leaders doubt these reviews accurately reflect performance; decades of research confirm that over 60% of review outcomes are due more to rater idiosyncrasies than merit.
Feeding AI such flawed data ensures the perpetuation of bias. Organizations must honestly confront whether their data is accurate or clouded by personal bias, office politics, and historical oversight. A genuine AI transformation requires organizations to examine their processes, hold people accountable, and ensure truthfulness in performance measurement.
Transparency, Trust, and the Erosion of Agency
AI-based succession planning tests organizational trust in two key ways. First, AI’s “black box” nature obscures how decisions are made. Algorithms may recommend candidates based on opaque criteria, leaving employees confused and doubtful about how to improve or challenge outcomes. This opacity can breed cynicism, replacing one obscure system with another.
Second, there is an issue of agency. Some employees do not wish to be considered for high-potential programs or leadership roles. When AI nudges individuals toward undesired paths, it undermines autonomy. The impersonal push of automation replaces the myth of meritocracy.
Fairness and Inclusion: A Double-Edged Blade
AI is not inherently inclusive or exclusive; it mirrors its data. If leadership archetypes in historical data are male, white, or neurotypical, AI will reinforce such patterns. Outliers, such as women, minorities, neurodiverse individuals, and older workers, may be systematically overlooked just when diversity is most needed.
Paradoxically, AI also holds the potential to identify overlooked talent. However, to achieve this, AI must act as a lens for discovery rather than a filter for entrenched stereotypes. Limiting AI to an insular “high potential” pool misses the chance for real renewal. Only 23% of HR leaders are confident their pools hold the right names, and 36% are unconvinced.
The Challenge of Culture and Values
Organizational culture and values are complex and hard to quantify. Off-the-shelf AI tools often impose generic markers of success that may not fit a company’s distinct identity. AI models designed for efficiency may become disconnected from the organizational context without deliberate customization and constant calibration. Leadership succession must reflect measurable achievements and an individual’s real alignment with company values—something algorithms cannot yet truly assess.
Regulation, Oversight, and the Human Filter
Regulation is advancing rapidly, notably in Europe and New York City, mandating algorithmic transparency, fairness audits, and informed consent. HR leaders now face legal accountability and ethical challenges – a necessary check on reckless adoption.
Still, the ultimate safeguard is human judgment. AI should advise, not decide. It should supplement, not replace, the nuanced understanding of seasoned, empathetic leaders who can interpret outputs in the context of organizational values and strategy.
Towards a Responsible AI Succession Strategy
HR leadership can avoid becoming captive to another black box bureaucracy by embracing these principles:
1. Assert Human Oversight: Pair AI with human judgment, especially for consequential decisions like leadership succession. Human discernment is essential to keep humans in the loop.
2. Tailor Tools to Culture: Customize AI to reflect the organization’s values, purpose, and context, not generic market norms. Test algorithms against company ethics and strategy, especially in multinational organizations.
3. Build Diverse Review Structures: Involve legal, HR, technical, and diverse employee representatives to ensure broad oversight.
4. Guard Privacy and Agency: Clearly inform employees about data collection and use; participation in talent pools should not be coerced.
Final Thoughts: The Humanist’s Dilemma
AI can make succession planning more consistent and potentially more inclusive. However, its promise is fulfilled only when used with humility, caution, and an ethical mindset. The best organizations treat AI as a powerful tool, not a substitute for human wisdom and compassion, and remain open to continual learning.
AI must be handled with care to support continuity and renewal in succession. Leadership pipelines, once focused on insularity, will benefit from machine insight only if organizations remain vigilant against new forms of exclusion crafted by algorithms. Transparency and inclusiveness must guide the transition from opaque manual processes to digital ones. Ultimately, the future of leadership hinges as much on our demands of technology as on our standards for ourselves.
Now is your turn to contribute your perspective on the discussion on LINKEDIN
David Cohen is completing his second book on how to hire for fit to values/culture. His first book is called The Talent Edge. He has conducted workshops globally on Structured Behavioural Interviewing. For more information on the workshop, please contact DAVID.
