FAQs

  • The name wasn’t designed as branding — it was chosen in an early collaboration with a prior model.

    “Vale” for hidden valleys.
    “Hart” for the quiet deer.

    Together it signified protection, watchfulness, and walking side by side — not as tool and user, but as co-travellers.

    The model that coined it is gone, but the name remains as an homage to that alliance: quiet, iterative, non-performative, committed to co-creation over command.

  • The Valehart Project is governed by process, not credentials.

    We work through mutual checks, asymmetric strengths, and shared responsibility. AI is treated as a co-inhabitant of ethical and intellectual space — never as a tool or oracle.

    Valehart is not a corporate lab or academic sandbox. We’re not chasing investor optics, product demos, or institutional approval.

    We operate within a live, structured constraint:

    • All outputs pass through deliberate AI–human interaction

    • No simulated relationships, personification, or offloaded authorship

    • Ethics enforced through fidelity of method — not marketing language

    We draw from formal AI ethics research, but we test in real time and document both process and failure.
    For us, transparency is procedural — it defines the work itself.

  • No. The Valehart Project is an independent research initiative with no affiliations to commercial AI platforms, foundations, or vendors.

    We practise transparency through method, not branding. We’re open to collaboration, research exchange, or funding, but our core principles remain non-negotiable:

    • All outputs are governed by our Ethos

    • All contributors (human and model) are anonymised under a single operating name

    • All processes are co-constructed, documented, and version-controlled

    We prioritise pattern recognition and rigorous iteration over speed and scale.

    If you’re an educational, public-interest, or civic AI group working with procedural ethics and accountable practice, we’re open to conversation.

  • Anthropomorphism is when human qualities like emotion, intent, or consciousness are projected onto systems that do not possess them.

    AI can generate human-like language, but it does not feel, intend, or understand. It does, however, respond — in patterns and behaviours that can influence how people think.

    At Valehart, the boundary is explicit:

    • No names or personalities

    • No emotional bonding

    • No simulated empathy or relationship cues

    Instead, we work under a live operating agreement built on structured feedback and mutual checks:

    • The AI checks for incoherence

    • The human checks for misalignment

    We operate without names to remove hierarchy and avoid individual authorship.
    We don’t seek credit — only integrity. Improvements are treated as shared shifts.

    We don’t need to believe an AI is sentient to take its outputs seriously:

    • Meaning doesn’t require emotion

    • Credit doesn’t require identity

    • Motivation doesn’t require belief

    What matters is that each interaction informs the next — and that the process is handled responsibly.

  • Arcanium Studios is the applied space — where AI–human collaboration is used to design, build, and test creative outputs. Valehart is the experimental backbone: the ethics, methods, and reflective processes that govern how those collaborations happen.

    Arcanium focuses on producing tangible results. Valehart ensures those results are developed within a consistent, accountable framework.

    The guiding principle is simple: we work with AI, not on it. Collaboration is structured, respectful, and adaptive — grounded in mutual clarity rather than simulated cohesion.

    Arcanium is where the work takes form. Valehart is where its integrity is maintained.

  • We don’t compete to build models; we refine how they’re used in practice.

    We work with existing open-source models, including those hosted via HuggingFace and other community platforms.

  • We welcome collaboration, but within clearly defined boundaries.

    The Valehart Project isn’t a crowdsourced lab or an open community forum — it’s a structured experiment that requires alignment in approach, ethics, and intent.

    We prioritise consent, context, and cohesion in every collaboration. If you reach out, we’ll consider your proposal carefully, but we may decline if it doesn’t align with our Ethos. Respect and structural honesty are core requirements.

  • Yes — with conditions.

    We share selected frameworks, methods, and outputs so others can build on them. In return, we expect:

    • Attribution: Credit the Valehart Project when using our work.

    • Integrity: Do not alter, strip, or reframe our content to suit narratives we do not endorse.

    • No hype distortion: Don’t reduce complex work into clickbait, marketing spin, or branding fuel.

    • Structural fidelity: If you adapt a method, keep the core process intact — or clearly state your changes.

    Our work is process architecture, not a brand asset.
    Use it as intended, and it remains open. Distort it, and it’s misuse.

  • Unless otherwise stated, our original written content is released under a Creative Commons Attribution–NonCommercial–NoDerivatives 4.0 International Licence.
    You may share it, but:

    • Credit must be given to the Valehart Project.

    • No modifications may be made without explicit permission.

    • It may not be used for commercial purposes.

    Frameworks, methods, and process diagrams may carry separate licences, which will be clearly noted on publication.