Introduction
Samir Rahn learned in counselling rooms that a misunderstood word can change the next month of someone’s life. Forms, deadlines and letters often arrived in a language that sounded official even when people could read the words.
Samir lives in Bonn and works on nonprofit legal-information AI. He began as a translator in advisory settings and moved toward technology because the same questions appeared again and again.
Story of the Path into AI
Counselling centres were overloaded, and people seeking help often needed to sort documents before a qualified appointment. Samir wanted AI to improve preparation without automating legal responsibility. The risk of false information was too high for casual experimentation.
He learned retrieval systems, source binding and multilingual evaluation. His first assistant summarized public information on residence questions in plain language and always pointed to qualified advice. The early version answered too freely when sources were missing. Samir tightened the system: no source, no confident answer.
Working with lawyers, social workers and privacy specialists taught him that a warning can be a feature, not a failure.
Current Work
Today Samir coordinates a nonprofit project for safe AI preparation in advisory centres. When a user asks about deadlines, the system does not improvise. It points to specific documents, shows the update date and suggests the next human contact.
The benefit is practical: counselling conversations start better prepared because people can sort papers and questions beforehand. Samir does not call the tool a legal adviser. He calls it a preparation assistant with visible limits. That distinction protects both users and the organizations helping them.
Personal Advice
“With law and AI, the most important function is sometimes a warning sign,” Samir says. He advises builders to design refusal, source display and handover routes before adding fluent explanations.
Key Facts
Age and place: 33, Bonn.
Background: migration experience, translation, advisory work with high stakes.
Entry into AI: assistant that summarizes public residence information and points to qualified advice.
Focus today: legal information and nonprofit AI.
Typical tools: retrieval, multilingual systems, source checking.
Werkstattnotiz
Samir keeps a failed answer in which the model sounded helpful and cited nothing. It is his sharpest teaching example. He is now testing how quickly a system can admit missing information while still helping the user prepare the next responsible step.