In this episode of The Mindful CIO, Nathalie Heynderickx is joined by Paul Cooper, a consultant and researcher in digital health, to explore the practical realities of responsible AI adoption.

Drawing on a diverse career spanning science, technology, and organisational change, Paul reflects on how generative AI differs from earlier approaches and what this means for leaders today. The discussion moves beyond tools to focus on how organisations can build trust, culture, and capability in a rapidly evolving environment.

A central theme is the distinction between structured and unstructured problem domains. While AI can perform reliably in closed, testable contexts such as coding, its application in business decision-making remains less predictable. This creates both opportunity and risk for CIOs and senior technology leaders.

The conversation also examines the current state of AI adoption. Many organisations are still experimenting, often without clear use cases or governance models. Paul outlines the importance of developing critical thinking skills, selecting appropriate use cases, and avoiding over-reliance on AI as a source of truth.

Governance is addressed through a practical lens, including tiered risk frameworks, the limitations of existing standards, and the role of boards in overseeing AI-related decisions. The discussion highlights the need to balance innovation with accountability, without constraining progress.

The episode concludes with reflections on organisational culture and capability. Leaders are encouraged to foster informed experimentation, support communities of practice, and develop the judgement required to use AI effectively.

This episode offers a thoughtful perspective for leaders seeking to implement AI in a responsible and sustainable way.

_____________________________________

Note: Although this episode is published in April 2026, it was recorded in December 2025. Paul refers to ChatGPT 5.2, which at the time he didn't rate highly for coding. His perspective has since evolved, and he now uses Codex 5.4 daily, alongside Claude.

_____________________________________

🎙️LISTEN & SUBSCRIBE TO OUR AUDIO PODCAST

✦ Website: ⁠⁠⁠⁠https://mindfulcio.community/podcast⁠⁠⁠⁠/

✦ Apple: ⁠⁠⁠⁠https://podcasts.apple.com/podcast/mindful-cio-podcast/id1765106205⁠⁠⁠⁠

✦ Spotify: ⁠⁠⁠⁠https://open.spotify.com/show/1gomcBWSDwdfoEZeZD7Usu_____________________________________🔗CONNECT WITH PAUL COOPER:

* Linkedin - https://www.linkedin.com/in/digitalfella/

*Longboardfella Consulting - https://longboardfella.com/

*Health Ability - http://www.healthability.org.au

* BioGrid Australia - https://www.biogrid.org.au/

* Australian Institute fo Digital Health - https://digitalhealth.org.au/_____________________________________📚 RESOURCES FROM PAUL COOPER

* A governance model for the application of AI in health care - https://academic.oup.com/jamia/article-abstract/27/3/491/5612169?login=false

* Reasoning Models don't always say what they think - https://www.anthropic.com/research/reasoning-models-dont-say-think

* Interpretability - Understanding how AI Models think - https://www.youtube.com/watch?v=fGKNUvivvnc