RA-01Research agenda — not results
Reproducible computational research records
A proposed programme for making computational claims inspectable from research question through data transformation, analysis environment, and reported conclusion.
Record SITE-REGISTER-01Reviewed 2026-08-29
A research-information prototype that states its limits first
BBQ demonstrates how a small research initiative could publish provenance-rich agendas, methods, and evidence records without inventing institutional status, people, studies, affiliations, or outcomes.
Public document index
RA / 03
Research agenda
Research agenda — not resultsProposed questions in reproducible computation, scientific provenance, and transparent AI-assisted research.
DP / 03
Demonstration projects
Demonstration project — no real studyWorked documentation structures using invented records, no participants, and no empirical outcomes.
SP / 03
Synthetic role profiles
Synthetic profile — not a real personNon-human role models that demonstrate responsibility fields without creating identities or credentials.
CA / 05
Cases register
synthetic + externalFictional quality-control scenarios separated from independently documented external cases.
ER / 08
External literature
External reference — not a BBQ outputReal scholarly records with original authorship, DOI, venue, source notes, and no BBQ output claim.
Institutional status at a glance
Every status value below is deliberately conservative. An absent legal or academic fact is shown as absent, not inferred from the domain, design, terminology, or subject matter.
| Status field | Declared value | Interpretation |
|---|---|---|
| University | No | No university status is claimed. |
| Accredited institution | No | No accreditation evidence exists for this initiative. |
| Degree-granting authority | No | The site offers no courses, credits, qualifications, or degrees. |
| Registered nonprofit or charity | Not claimed | Independent and non-commercial describes intent, not legal status. |
| Relationship to Anthropic | None | No affiliation, sponsorship, endorsement, authorization, or partnership is claimed. |
| Real personnel directory | No | Role profiles are synthetic interface demonstrations. |
| Real internal research outputs | None claimed | Research pages describe agendas and demonstrations, not completed studies. |
| External literature | Clearly separated | External authorship, affiliation, venue, DOI, and rights are preserved. |
Current research agenda
RA-02Research agenda — not results
Scientific metadata and claim provenance
A proposed study of how publication metadata, primary sources, correction status, and local interpretation can be separated in a durable evidence register.
RA-03Research agenda — not results
Transparent documentation for AI-assisted research
A proposed framework for recording where generative systems enter a research workflow, what evidence they can access, and which decisions remain human responsibilities.
How records are classified
| Class | What it may contain | What it cannot establish |
|---|---|---|
| Research agenda | Proposed questions and methods | Not a completed study or result |
| Synthetic profile | A fictional role model for interface demonstration | Not a real person, appointment, credential, or affiliation |
| Demonstration project | A worked documentation pattern | No real participants, funding, partner, or outcome |
| Synthetic case | A fictional quality-control scenario | Not a client, incident, or impact claim |
| Synthetic data | Deterministically authored rows | No real subject, organization, or measurement |
| External reference | An independently published scholarly record | Not a BBQ output or relationship |
| External documented case | A source-grounded account of external work | No BBQ role identified in cited sources; no relationship claimed |
| Site policy / draft policy | Rules applied to the website or a labelled governance model | Not accreditation or legal-entity governance |
Selected external reading
These works provide background vocabulary for the agendas. They are independently authored and published; selection for this catalogue is not authorship, endorsement, collaboration, or institutional affiliation.
External reference — not a BBQ outputREF-DATASHEETS-DATASETS
Datasheets for datasets
Communications of the ACM (2021). DOI: 10.1145/3458723
The framework asks dataset creators to document why a dataset exists, what it contains, how it was collected and processed, where it should and should not be used, how it is distributed, and how it will be maintained. It gives downstream users context that files alone cannot provide, while remaining documentation rather than a fairness, legality, or quality certificate.
- Publication status
- active
- Affiliation field
- partially present
- Last checked
- 2026-08-29
External reference — not a BBQ outputREF-PRISMA-2020
The PRISMA 2020 statement: an updated guideline for reporting systematic reviews
BMJ (2021). DOI: 10.1136/bmj.n71
PRISMA 2020 updates the reporting checklist and flow-diagram framework for systematic reviews. Its relevance here is structural: a reporting guideline can make methods and selection decisions inspectable, but completing a checklist is not by itself proof that the review question, included evidence, analysis, or conclusion is correct.
- Publication status
- active
- Affiliation field
- not present in source
- Last checked
- 2026-08-29
External reference — not a BBQ outputREF-STOCHASTIC-PARROTS
On the Dangers of Stochastic Parrots
Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (2021). DOI: 10.1145/3442188.3445922
The paper examines risks associated with increasingly large language models, including resource costs, undocumented training data, encoded harms, and misleading interpretations of generated text. It is relevant to disclosure design because it asks what developers and researchers should know before treating fluent output as meaningful evidence.
- Publication status
- active
- Affiliation field
- present
- Last checked
- 2026-08-29