AI as part of the interaction

Parallaxe is designed around a conversation between two people with opposing political perspectives, mediated in real time by artificial intelligence.

The system can reformulate emotionally charged statements, ask clarifying questions, translate between languages and introduce source-based context when factual claims can reasonably be verified.

Rather than functioning as a separate chatbot, the AI becomes part of the structure of the conversation itself.

From responsive systems to agent systems

One of the most interesting developments in AI is the shift from systems that simply generate responses toward systems that can observe context, make decisions and perform different roles within a larger workflow.

Parallaxe works with this logic through multiple specialised AI processes that contribute to mediation, translation, clarification and factual verification.

For me, this makes agent systems particularly interesting from a design perspective. As AI gains more autonomy, interaction design increasingly needs to address what the system is allowed to do, when it should intervene and how users remain aware of its influence.

Designing for agency and control

Interactive systems are often evaluated by how seamlessly they respond to users. With AI, seamlessness is not always the only goal.

When a system mediates communication between people, users also need to understand when an intervention has taken place and what role the AI is playing within the exchange.

This creates a balance between fluid interaction and meaningful human control. The system should be helpful enough to support the conversation without becoming an invisible authority within it.

Human-AI interaction beyond individual users

Many AI interfaces are designed around a relationship between one person and one artificial system.

Parallaxe introduces a different configuration: AI becomes an intermediary between multiple people.

This changes the design problem. The system is not only responding to an individual user, but influencing a social interaction in which participants may have different goals, assumptions and expectations.

For me, these multi-user forms of human-AI interaction are an important part of thinking about future agent systems.

Bringing Parallaxe into new contexts

Presenting Parallaxe at gamescom introduced the project to a community already familiar with interactive technologies, digital environments and emerging forms of interface design.

It also placed political dialogue and responsible AI alongside broader conversations around interaction, agency and technological experimentation.

Moving the project across research, cultural and technology contexts continues to be valuable for understanding how different audiences interpret the role of AI and which questions become visible when the same system is encountered from different perspectives.