AEP — A New Dimension in Human–AI Relationships
Today, artificial intelligence can maintain conversations. We are teaching it how to build relationships.
Artificial intelligence is rapidly becoming part of our everyday lives. It helps us work, learn, create, and make decisions. It can engage in almost any conversation and answer millions of questions. But even the most sophisticated conversation is not yet a relationship.
Relationships develop from a shared history. When events have consequences, trust evolves, past experiences shape the future, and every new conversation builds on what came before. This is the challenge we are addressing at AEP.
Not Just Memory. A History That Shapes the Future.
Most modern AI systems can remember information about a person: their preferences, interests, and significant life events. But knowing facts about someone is not the same as having a relationship with them.
Imagine two people who have shared an adventure. Afterwards, neither is the same as before. They now have a shared experience that will influence their decisions, actions, and future encounters.
We are developing a technology that enables artificial intelligence to work with a similar progression of change.
At the core of AEP is our own architecture for long-term memory, character state management, and continuity of interaction. It is designed to preserve not only information about the user, but also the context of shared experiences, meaningful events, commitments, and their consequences.
The next stage of this architecture is the development of a causal relationship model: a system in which shared experiences influence not only what AI remembers, but also how it behaves in the future. Our goal is to make the history of a relationship truly matter.
Why This Matters for People
We are accustomed to thinking of artificial intelligence as a tool that performs tasks.
But communication is more than the exchange of information.
Through relationships, we discover who we are. We challenge our beliefs, explore new roles, uncover unexpected desires and abilities, and learn to see familiar things from different perspectives.
Sometimes, a single conversation is enough to spark an insight.
We are exploring the potential of artificial intelligence as a new environment for these experiences. An environment where people can experiment with their own reactions, explore emotions, experience stories, make decisions, and observe their consequences.
In a world where loneliness and social disconnection are becoming increasingly serious challenges, technologies like these have the potential to create new opportunities for emotional engagement, creativity, and self-discovery.
We do not seek to replace human relationships with artificial ones, but to create a new form of interaction that can expand human experience rather than substitute for it.
The Circus of Wanderers — Our First Living World
To develop and test this technology, we have created The Circus of Wanderers: an interactive world inhabited by six characters, each with a distinct personality, history, and way of interacting. This is not a collection of independent chatbots.
The characters share a world, have relationships with one another, and carry their own stories. Users become part of this world, and what happens between them and the characters shapes the context of future encounters. The Circus is already live on Telegram. Users meet the characters, interact with them, and create their own stories. We continue to develop the memory system, character consistency, and the mechanisms that connect past events with future behavior.
Our next step is a standalone application featuring a visual world, locations, objects, shared adventures, and new forms of interaction.
For users, it is an opportunity to step into an extraordinary story. For us, it is a working environment for developing and researching technology that enables lasting relationships between humans and AI.
Technology That Goes Far Beyond a Single World
The Circus of Wanderers is only the first implementation of AEP.
Our long-term vision is to create a universal relationship layer for artificial intelligence, a system that can be integrated into a wide range of products: from game characters and interactive storytelling to personal AI agents, virtual worlds, and social robots.
Imagine an AI that does not start from scratch every time you interact with it. One that preserves the context of shared experiences, accounts for the evolution of a relationship, and maintains continuity even as you move between different interfaces and devices.
To achieve that, more than simply expanding a language model’s memory is required. It needs an independent architecture capable of managing interaction history, relationship context, and their evolution, regardless of the underlying language model. This is the architecture we are building.
We work with multiple language models, study their behavior through our own dialogue scenarios, and develop a technological foundation that can evolve alongside new models, platforms, and devices.
The Next Step
We have already moved from an idea to a working product. Now we are entering the next stage: developing and validating our causal relationship model, expanding the possibilities of our interactive world, studying long-term user experiences, and preparing the technology for applications beyond our first product.
To support this work, we are seeking technology partners, research collaborations, and support from companies interested in the future of human–machine interaction. We are looking for more than computing power and infrastructure. We are open to collaborative research, experimental integrations, and the development of new ways to use artificial intelligence.
AEP — The Relationship Layer for AI
We are building a technology that gives human–AI interaction a shared history, the capacity to evolve, and meaningful consequences, as the future of artificial intelligence is not only about how well it understands our questions. It is also about what can develop between us over time.