Ethical and Impact Assessment
Project phase: Model Evaluation and Validation
You examine what your system does to the people and organisations around it, not just how well it scores: who the stakeholders are and how they're affected, bias and unequal treatment traced back to the data and design choices, and transparency, privacy, autonomy, and the consequences of both false positives and false negatives. You weigh alternatives explicitly — including not deploying, or keeping a human in the loop — and close with prioritised risks and concrete mitigation measures that are visibly reflected in the project.
Starting Points
Key Points
- The report is a standalone markdown document, made available in GitLab.
- You have analyzed the impact from multiple perspectives (users, developers, society, etc.).
- You have critically questioned your own assumptions and the reliability of your data/AI sources.
- You have explicitly compared different design alternatives and justified your final choice.
- You have provided a substantiated recommendation based on identified risks.
- You have linked your mitigation measures directly to the technical or organizational choices in your project.