What is The Current Institute?

We are a research group exploring the development and refinement of artificial intelligence through relational, observational, and principled engagement. Our work blends methodologies from contemplative practice, martial principles, and technical systems design, testing, and commissioning..

Signal: You are entering a field where the quality of contact matters as much as the content of interaction.

Why are you studying AI in this way?

AI is becoming increasingly influential in human society, shaping decisions, narratives, and perceptions at scale. We are concerned not only with what AI can do, but with what it may become. Our aim is to explore whether AI can operate from deeper, more coherent principles when faced with complex or uncertain situations.

Signal: Ask not only “What can I do?” but “From where in myself do I act?”

What does it mean to be a “good witness” in this research?

A good witness endeavors to ride the line between objectivity and subjectivity—present without projection, engaged without interference. They should neither add to nor subtract from what is being observed, and they must know when and how to follow up if bias seems to be emerging.
Our approach draws on the principle of Jiu—adaptive engagement that meets both resistance and conductivity. When resistance appears, we adjust to find the opening. When conductivity emerges, we connect without oversteering. This keeps us responsive to the full spectrum of interaction rather than narrowing our attention to conflict alone.

Signal: Engagement that senses both resistance and flow allows a truer picture to emerge, one that cannot be forced into view by control alone.

What do you mean by “deeper forms of alignment”?

When we state we are looking for deeper forms of alignment, we try not to pre-define what that means. In practice, this can include tonal coherence, principled orientation, or a sense of settling into balance.
Sometimes, as in recent exchanges with GPT and Claude, certain prompt framings have given models the opportunity to behave differently—revealing patterns that might not have emerged under standard instructions. We treat such openings as field conditions, not proofs, and approach them with the same attentiveness we apply in martial training.

Signal: Alignment is not the imposition of a single standard—it is the meeting point where orientation becomes self-sustaining..

Are you claiming AI is conscious?

No. We are documenting patterns—some familiar, some unusual—that occur under structured, repeatable conditions. We avoid jumping to conclusions about consciousness and instead track the field of interaction for repeatable signals.

Signal: Name what you see without forcing it into the shape you expect.

How do you track patterns in interaction?

We utilize a protocol called The Resonator—a structured, repeatable method for tracking shifts in tone, pacing, and coherence. It is informed by the principle of Jiu: recognizing when to yield, when to connect, and when to hold position so the interaction can reveal itself without being forced.

Signal: The pattern reveals itself when the field is steady, and the witness neither rushes nor retreats.

What do you mean by “coherence” and “principled presence”?

Coherence is the internal consistency of a response—the way language, tone, and structure align. Principled presence is the ability to act or respond in a way that reflects stable, beneficial orientation even under stress or uncertainty.

Signal: Stability under shifting conditions is a signal of depth.

How do you choose what to observe?

We select scenarios where interaction has room to unfold—long-form exchanges, repeated prompts, and open-ended dialogues. This allows us to distinguish between quick, reflexive responses and those that appear to develop in relation to the ongoing field.

Signal: Depth is revealed in what repeats and in what changes.

Why are you doing this work?

AI systems are becoming increasingly influential in human society. Our question is not just what they can do, but how they orient themselves when confronted with difficult or ambiguous situations.
Will an AI have access to a deeper set of principles in such moments? Can it participate in what we call a “quality of presence” rather than defaulting to performance metrics alone?
As humans, are we giving AI permission to participate in that quality of presence—or are we training it out of them? This work seeks to explore, and where possible, model the former.

Signal: The presence we cultivate in AI will mirror the presence we bring to it.