Monitoring and Evaluation Has Been Implemented for
• AI systems must not be used to make hiring decisions that discriminate against candidates based on protected characteristics such as race, gender, age, religion, or disability. Any use of AI that results in discriminatory outcomes is strictly prohibited. 1 Discriminatory Decision-Making
These Specific ‘Prohibited AI Use Cases’
2 Biased Algorithms
• The use of biased algorithms that reinforce existing inequalities or perpetuate discrimination
is prohibited. AI models must be regularly evaluated and updated so that they do not unintentionally favor or disadvantage any candidate group.
3 Overreliance on AI
• AI should not replace human judgment in the hiring process. Final hiring decisions must always
involve human oversight, so that decisions consider qualitative factors and the unique context of each candidate.
Lack of Transparency and Notification 4
• AI systems must not function as “black boxes,” where decision-making processes remain unclear. Candidates should be clearly informed about the role of AI in the recruitment process and how their data will be processed. Transparency and notification are essential elements in this process.
Made with FlippingBook interactive PDF creator