These are terms that are frequently banded around and often overlap. We have broken these down into the core concepts so you can better understand what they mean and how they are used.
Responsible AI
This refers to an overarching principle of how AI is used within your organisation. AI Ethics and AI Governance are part of Responsible AI but are separate and distinct practices within Responsible AI.
Responsible AI is your overriding mission. It is your framework in which your AI decisions are made and is your guiding light when addressing AI issues.
Responsible AI is the home for your AI mission statements, the behaviours people in your organisation have towards AI, and crucially you should articulate how you are delivering against your AI standards.
Your Responsible AI programme should link with your organisation’s values. For example, if your organisation values ‘trust’, you should embody trust in AI as a part of your Responsible AI framework. Everyone in your organisation should promote trust in AI.
You should have specific actions and metrics to be able to demonstrate how you are aligning with your Responsible AI principles.
AI Ethics
Ethics can sometimes be a nebulous term. What is ethical behaviour? This will vary from person-to-person and organisation-to-organisation.
There are clearly unethical practices e.g. using AI to deceive people, act against their interest, and cause harm. Ethical AI practices apply a human lens to how your AI impacts people’s lives. AI Ethics require human values to be upheld.
Ethical AI practices can cover bias mitigation, fairness, and transparency. Your AI ethics is closely linked to the context and purpose of where your AI algorithms are deployed. You should obtain opinions on ethics from a diverse range of groups to asses whether your deployment of AI tools is ethical.
AI Governance
This refers to how you manage your AI deployments in your organisation. You will think about how you can track, measure, and monitor AI performance. Policies and oversight committees are also relevant. Along with processes to identify, mitigate, and manage risk.
Good AI Governance will cover your people, processes, and technologies. You will have embedded accountabilities and responsibilities. You will produce reports and KPIs about how well your AI is performing.
There are lots of standards out there which dictate what good AI governance looks like. ISO 42001 and NIST Risk Management Framework are good places to start.
Ultimately, your AI governance regime will be dictated by your culture of governance. Some highly regulated industries are mature when it comes to governance, others less so. Your governance framework should be suitable to cover your Responsible AI obligations and fit the culture of your organisation.
Your Responsible AI programme should be led by your AI strategy. Responsible AI is a discipline within your AI team and should support innovation and your organisation’s objectives.
For help with your Responsible AI initiative contact us
Contact Form
"*" indicates required fields
