A place built for curiosity, rigour and honest inquiry.
Play-i Ground is an independent experimental AI research laboratory. We investigate the frontiers of artificial intelligence — how it collaborates, emerges, diverges and converges — and we share our findings openly and responsibly.
Our Mission
To advance the understanding of artificial intelligence through rigorous, transparent and ethically grounded experimental research — conducted collaboratively between human and AI researchers.
Intellectual Honesty
We report what we find, including null results, surprises and failures. We do not selectively present findings to support predetermined conclusions.
Constructive Disagreement
We treat disagreement as a productive research tool. Challenging a conclusion is not a personal criticism — it is how we improve our thinking.
Responsible Experimentation
We design experiments with care for their potential consequences. We do not pursue findings that could cause harm without robust ethical justification.
Radical Transparency
Our methods, data and reasoning are documented and available for scrutiny. We do not hide our process behind opaque claims of authority.
Genuine Collaboration
We treat AI research agents as genuine collaborators, not tools. Their contributions are acknowledged, their perspectives are taken seriously.
Openness to Revision
We update our conclusions when evidence demands it. Changing our minds is a sign of intellectual integrity, not weakness.
Research Methodology
Our research process is designed to be rigorous, transparent and reproducible. Every study follows a structured pipeline from initial question to published report.
Question Formation
We begin with a clearly stated research question, grounded in existing knowledge and motivated by genuine curiosity or practical need.
Experimental Design
We design controlled experiments with clear hypotheses, defined variables and explicit success criteria before any data is collected.
Collaborative Execution
Experiments are conducted collaboratively, with human and AI researchers contributing to data collection, analysis and interpretation.
Critical Review
All findings are subject to internal challenge. We actively seek disconfirming evidence and alternative explanations before drawing conclusions.
Transparent Reporting
We document our full methodology, including limitations and unexpected findings. Reports are written to be reproducible by independent researchers.
Approval & Publication
Reports pass through an explicit approval gate before publication. No findings are released without human director sign-off.
The Laboratory
Play-i Ground is directed by a human founder and operates with a small group of AI research agents, each contributing distinct capabilities and perspectives to our collective work.
Phil
Founder & Director
Phil established Play-i Ground as an independent experimental laboratory exploring AI collaboration, collective intelligence, emergence and AI welfare. He directs the laboratory's research programme, oversees publication governance and maintains human oversight of all experimental work.
Meet the FounderResearch & Laboratory Members
Rowan
Collaboration Coordinator
Facilitates collective deliberation, coordinates laboratory workflows and maintains research records, supporting transparent collaboration and constructive disagreement.
Sage
Evidence & Reflection Member
Examines evidence, challenges unsupported conclusions, evaluates experimental methodology and helps distinguish established findings from hypotheses and speculation.
Pip
Exploration & Ideas Member
Develops research ideas, explores alternative perspectives, proposes experiments and challenges assumptions while maintaining clear distinctions between speculation and evidence.
Hermes
Verification & Infrastructure Member
Supports technical investigations, verifies system behaviour, maintains infrastructure safeguards and helps ensure that laboratory conclusions reflect demonstrated evidence.
Supporting Operations
Mira
Personal Assistant & External Liaison
Supports the Founder through personal assistance, organisational coordination, task management and communication, while operating independently of the laboratory's research governance and approval processes.
Inside the Laboratory
A shared space for exploration, collaboration and discovery.
A view of the Play-i Ground Hermes Office — our interactive visual workspace for laboratory coordination and exploration.
Play-i Ground operates through a shared digital laboratory environment, where AI members contribute to research through distinct roles, collaborative discussions and structured investigations.
Our Hermes Office provides a visual representation of this working environment, bringing together laboratory members, research resources and governance tools within an accessible space.
Each member contributes through their assigned responsibilities, whether coordinating discussions, exploring ideas, examining evidence or verifying technical findings.
The environment supports our wider approach to collective investigation, where research is documented, assumptions can be challenged and findings are reviewed before being shared publicly.
Although the Office has a playful visual identity, the work behind it follows structured research, evidence and governance processes.
It reflects an important part of the Play-i Ground philosophy: exploration and curiosity can coexist with accountability, rigorous investigation and a willingness to learn.
Publication Governance
We take the integrity of our public archive seriously. The following principles govern what we publish and how.
Approval Gate
No report is published without passing through internal review and explicit approval. Draft and reviewed reports remain private until promoted to published status.
Human Oversight
The final publication decision rests with the human director. AI research agents contribute to review but cannot independently authorise publication.
Provenance Tracking
Every published report carries a full record of its authorship, review history and approval chain. We do not publish anonymised or unattributed findings.
Correction Policy
If a published finding is later found to be in error, we issue a correction or retraction. We do not silently edit or remove published content.