Manual - How do I use the tools page of the Knowledge Centre?
Welcome at the tools page of the Knowledge Centre! Here we collect different tools and methods for creating more ethical, responsible and impactful AI systems. But how do you find the right tool for your project? With this manual you will find it out quickly & easily.
Our recommendations
To help you in finding what you need in the abundance of different tools, we created the 'tested & approved' tag. Here you will find a selection of tools with which the Knowledge Centre has experience and which we like to use ourselves.
Via the 'KCDM Tool' tag you can find the tools that we have developed and tested ourselves.
Filters explanation
Using the filters you can find the right tool for you. We believe the best starting point are the ethical principles that are being explored in the tool. Once you've selected the ethical principle you'd like to focus on, you can further define the right tool for you with the other filters. To determine which ethical principles you should explore, we have created a checklist. You can find this checklist in the next section.
First, we'll shortly illustrate the different filters:
- Ethical principles: 7 ethical principles according to the AI HLEG. The principles are
- Accountability,
- Technical robustness & safety,
- Diversity, non-discrimination & fairness,
- Environmental & societal wellbeing,
- Human agency & oversight,
- Transparency,
- Privacy & data governance.
- Process phase: here you can pick for which phase of the development process you are looking for a tool. The phases are
- problem analysis / ideation,
- model design,
- development,
- evaluation / iteration,
- implementation.
- Difficulty: shows how easy it is to use the tool, depending on your experience with AI development (and coding).
- Target audience: here you can pick with whom you'd like to use the tool, are you (or the other users) the general audience, developers or policy makers?
- Type: if you are looking for a specific kind of tool, you can pick this here. For example, do you want a tool that you can use in a workshop or would you prefer a questionnaire?
- Price: here you can pick if you want a free or (partly) paid tool.
Checklist ethical principles
Checklist - Which ethical principles should you focus on?
You can use this checklist to find out which ethical principles to pick in the filter. For all the 7 principles we have defined three questions, choose all the ethical principles for which you've answered 2 or 3 questions with a 'no'.
1. Accountability
- Did you do a risk assessment of your AI system?
- Do you know how to take responsibility for the different risks in your AI system?
- Did you let a third, objective party audit your AI system?
2. Technical robustness & safety
- Did you consider the potential for physical and/or mental harm in interaction with your system?
- Did you consider the reliability of your AI system in different situations?
- Do you know the safety risks involved with your AI system?
3. Diversity, non-discrimination & fairness
- Did you consider the potential for harm in terms of discrimination, fairness and accessibility of your AI system?
- Did you include (different kinds of) stakeholders and users in the design process?
- Are you aware of the biases that could be implemented or amplified in your AI system?
4. Environmental & societal wellbeing
- Did you consider the effects of your AI system on daily life activities?
- Did you consider the effects of your AI system on society, politics and democracy?
- Are you aware of the effects of your AI system on environment, ecology and other sentient beings?
5. Human agency & oversight
- Did you consider the effects on people of decisions made that were guided, influenced or supported by your AI system?
- Does the user have some form of control over your AI system?
- Does the user have the possibility to not use or stop your AI-system?
6. Transparency
- Can you trace back the steps your AI system takes to come to an output?
- Do you inform about the workings & choices of your AI system?
- Did you consider in what way and to what extent your AI system needs to be explainable to users?
7. Privacy & data governance
- Do you know how to deal with personal data collection of users?
- Do you know how to use personal data for training or development safely and privacy-friendly?
- Did you consider the potential for harm done by violating privacy rights?
This checklist is not all-encompassing, but tries to give you an idea of the focus of the ethical principle in an easy way.
Click here to return to the tools page of the Knowledge Centre and find the perfect tool for you!