RESPONSIBLE AND ACCOUNTABLE AI
Despite their power and potential, AI tools are still just mathematical models. What happens when an AI makes a decision with a negative impact on an individual or organization? The model itself cannot face any real consequence. Instead, accountability for AI and responsibility for deciding when and how it should be used is a uniquely human ethical task.
Accountability can be seen as a feature of the AI system, a determination of individual or group responsibility, and a quality of the social and technological systems in which the cognitive tools function. For AI, accountability means that not only can the system explain its decisions, the stakeholders who develop and use the system can also explain its decisions, their own decisions, and understand that they are accountable for those decisions.
A related but distinct dimension is responsibility, wherein stakeholders ask the essential questions of not just whether they can develop and deploy a given AI tool but whether they should do so. Determining if an AI application is a responsible choice is a whole-of-enterprise activity, involving all stakeholders from the executive suite to the most junior data scientist. It demands a shared sense of accountability and an acknowledgement that when working with AI, there is a necessary standard of care.
Is your organization’s AI Responsible and Accountable AI?
- Is it clear who is accountable for AI and related outputs? Who is monitoring the AI and how frequently?
- How will you communicate if an AI operates outside of its function and causes a problem? Does your organization have an established protocol to follow?
- What are the consequences for the accountable parties when AI causes harm, damage, or violates regulations or laws?
- Does the potential good that a model can achieve outweigh the potential negative outcomes it could cause? Has your organization asked, is this worth it?
- Are there laws or regulations governing how models can be used given concerns over environmental protection, social equity, data protection, and public safety and wellbeing?
[elementor-template id=”4059″]