Beyond Human: Your Top Questions Answered on AI and the Future of Leadership

Question: What are the privacy and regulatory implications of coaching with AI? How will the insights be shared and secured?

Answer: Different AI applications will have different privacy policies and security standards. It’s very important to review those standards. Considerations include how the data is stored, who has access to that data, and if the LLM is training on that data.

For sharing insights from AI coaching within organizations, establish clear protocols that respect confidentiality while enabling necessary knowledge transfer. Consider creating anonymized summary reports that highlight patterns and key learnings without revealing personal details. Implement role-based access controls so that only appropriate stakeholders can view specific levels of coaching data. Always be transparent with coaching participants about exactly how their information will be used and shared.

Question: How can HR leaders help skeptical stakeholders — whether leaders or coaches — see the value of AI?

Answer: It’s important to help these stakeholders understand that AI is not just a productivity tool, but actually an intelligent assistant that can enhance human growth and performance. Sharing empirical studies on this topic can provide evidence, such as the research by Dr. Nicky Terblanche and colleagues published in PLOS ONE which demonstrated that AI coaching can effectively support behavior change and performance improvement. But the best way for them to understand the value would be through experiencing it. Therefore, providing a demo of an AI coach can be really helpful in opening the eyes of a skeptical stakeholder.

In addition, HR leaders should lead by example! Share anecdotes and use cases of where you used AI successfully, and where not.

Question: To what extent can getting better at the Ideate stage of the 3i Framework transfer to improving communication skills with humans?

Answer: The 3i Framework is the logical sequence for communicating with AI, but it also parallels how great communication works with humans. The more two or more people ideate, the higher quality of output, especially if the humans are engaged in designing something or problem solving.

When we get better at ideation with AI, we develop crucial skills that transfer directly to human communication: the ability to articulate ideas clearly, ask thought-provoking questions, build upon suggestions, and synthesize diverse concepts. This process trains us to be more open-minded, recognize different perspectives, and avoid getting stuck in conventional thinking patterns. These same skills are invaluable when brainstorming or problem-solving with colleagues, as they foster more productive and innovative discussions.

Question: With innovation curve speeding up, and requests to leaders increasing, even with the help of AI what are your thoughts on ensuring mental wellbeing and avoid burnouts for leaders?

Answer: Based on our research, Resilience is one of the top AI leadership skills. The reason is precisely because there is so much change, disruption, and demands on the leader. Leaders need to approach change with urgency and energy, but pace themselves and ensure they are setting boundaries, engaged in self-care, and leading the pace of change rather than adapting to an ever increasing velocity.

Crucially, to support this resilience, leaders should intentionally use AI to reduce cognitive load rather than simply adding more capacity for work. This means leveraging AI for routine decision-making and information processing, while preserving human energy for strategic thinking and relationship building. Organizations should also establish clear expectations around response times and availability, and create cultural norms that respect recovery time. Regular reflection on priorities and purpose helps leaders stay focused on what truly matters rather than being pulled in multiple directions by accelerating demands.

Question: I am experimenting a lot with ChatGPT. What other platforms would you suggest we should start to experiment with as well?

Answer: The four main LLMs we recommend at AIIR are ChatGPT, Claude, Gemini, and Grok. Each offers unique strengths: ChatGPT is versatile and widely accessible; Claude excels at nuanced communication and detailed instructions; Gemini integrates well with Google’s ecosystem and can process both text and images together (meaning you can show it visuals and ask questions about them); and Grok provides a more experimental approach with its ability to take risks and handle unconventional requests. Experimenting across these platforms can help you identify which tool works best for specific leadership development and coaching scenarios.

Question: How do we ensure AI doesn’t reinforce bias or bad behavior in leadership development?

AI is a tool and multiplier for the initiator, which is the human. If the human brings significant biases, AI can reinforce these biases. However, well designed AI coaches take this into account by asking probing questions and being more cautious in giving advice or guidance that may confirm the users biases.

LLMs have come a long way since their inception in late 2022 in terms of their awareness of bias and how to manage bias. Remember that the output of a generative AI is to a large extent dependent on the quality of the prompt provided. The person who prompts the LLM has a certain measure of control about how bias and AI output may influence leadership development, therefore make sure the input (prompts) into the LLM is supported by solid theoretical underpinnings.

Question: Why trust this and why trust you?

Answer: Trust can be granted A priori as a function of reputation, expertise, and credibility. However, trust grows over time as someone establishes consistency in the expectations being met and reliability is established. To not trust is completely fine, but to not experiment or take calculated risks comes with its own consequences.

In the context of AI coaching, we demonstrate trustworthiness through transparency about our methods, limitations, and data practices. We encourage organizations to start with low-stakes applications, measure results against established leadership development benchmarks, and gradually expand usage as confidence builds. Our systems are designed to complement human judgment rather than replace it, creating accountability through human oversight of AI recommendations and regular evaluation of outcomes.