Predictive hiring is the data-driven prediction of an application's success based on historic hiring data and performance metrics. Machine-learning models learn from patterns in past hiring trajectories and estimate the probability that a new application would result in a successful employee.
Benefits are faster selection, data-supported decisions and better comparability. Risks are significant: learning from historic data potentially amplifies historic biases. If men were promoted predominantly in the past, the model may score women lower – a classic bias effect.
The EU AI Act classifies predictive hiring as a high-risk application. Required are risk management, data governance, transparency, human oversight, accuracy and robustness measurements and concrete discrimination audits. Germany's AGG also remains fully applicable.
Lunigi avoids classical predictive hiring and focuses on semantic matching between profile and role requirements – with full decision authority retained by candidates.