Understanding Responsible AI – Implementing ethical and privacy Consideration on using AI Models
Responsible AI emerged in this session as a question of judgement rather than technology. As AI systems increasingly influence operational decisions, panellists argued that trust must be earned through transparency, explainability, and clearly defined accountability, particularly in industries where safety, compliance, and reliability are inseparable.
The discussion highlighted the widening gap between everyday use of generative tools and their deployment in industrial environments. While AI can generate recommendations at speed, participants stressed that decisions without explainable reasoning are effectively unusable in high risk contexts. Explainability was positioned not as an optional feature, but as a prerequisite for adoption, enabling operators to understand, challenge, and ultimately trust AI driven outputs.
Governance was framed as an extension of existing risk and safety management practices rather than a new parallel structure. Panellists explored how AI use cases can be classified by criticality, with levels of testing, oversight, and human intervention scaled accordingly. This approach was presented as a practical way to manage evolving risks without stalling innovation.
Ethical and privacy considerations were woven throughout the conversation. Data ownership, lineage, minimisation, and access controls were identified as essential safeguards, particularly in joint venture environments and highly regulated settings. The panel concluded by emphasising that responsible AI is fundamentally a cultural challenge, shaped by organisational behaviour and leadership choices as much as by technical controls.
Speakers
- Julie Thyne , Global Improvement Director, Dow
- Bernardo Braunstein , Manager, Operational Technology, OT Data, Noble Corporation
- Harvinder Singh , Platform & Services Manager, Knowledge Management, bp
Moderator
