What the development focuses on

AI mediation for polarised dialogue
Parallaxe uses AI as an active mediation layer between participants. Rather than generating arguments or deciding which position is correct, the system supports structured dialogue by reducing unnecessary escalation, identifying misunderstandings, and helping participants engage with each other's underlying perspectives.

Real-time reformulation and adaptive follow-ups
The system processes conversational input and can reformulate statements into clearer, more neutral language while preserving their intended meaning. Adaptive follow-up questions help surface assumptions, clarify ambiguous terms, and identify the actual points of disagreement between participants.

Multilingual AI interaction
Parallaxe is designed for conversations across languages. Translation is integrated into the interaction flow, allowing participants to communicate in their preferred language while maintaining a shared mediated conversation.

Source-based verification and AI fact-checking
For claims that can be objectively verified, the system can provide source-based context and supporting evidence. The challenge is not simply retrieving information, but designing when verification should be triggered, how uncertainty is communicated, and how sources are presented without disrupting the conversation.

From prototype to robust AI system
The collaboration with Fraunhofer HHI strengthens the technical development of Parallaxe, with a focus on system architecture, reliable AI behaviour, interaction logic, and deployment in public-facing contexts. The goal is to turn an experimental interaction concept into a scalable system that can be tested in cultural, research, and institutional environments.