Results for ""
| Model | Strengths | Limitations | Recommended Context |
|---|---|---|---|
| None | No cost | Nobody has organisation-wide picture of responsible AI practices | Organisations aiming tactical AI adoption |
| Advisory | Does not disturb power structures or processes | Can become a ‘talking shop’ - finding problems rather than solutions as the onus to act remains with company executives | Organisations with low AI maturity |
| Governance | Assured enforcement, least risk | Can become a barrier to adoption, absence of safe harbour could deter executives from assuming role in Ethics Committees | Organisations with lowrisk AI propositions |
| Multi-Instance, one for each solution | Functional excellence and focus | Complexity, high cost | Large organizations with federated operating models and accountability structures |
| Mediation | Problem solving, continuously builds expertise | New organisational construct - organisational change management (OCM) and development of new skills may ben needed | Strategic AI adoption |
ADDITIONAL RECOMMENDATIONS FOR MODELLING ETHICS COMMITTEES
- Demonstrate multidisciplinary expertise in technology and ethics, diversity and inclusion, and enterprise risk management
- Be extended organisational protections for members to raise objections and concerns openly
- Have at least one external member who is not subject to organisational pressures or goals
- Be obliged to maintain operational transparency, while allowing members to keep certain findings sufficiently confidential to protect them