1. THE PROBLEM

The Human Participant Protection Program (HP3) addresses a practical gap faced by independent researchers, small nonprofits, community research groups, and investigators working outside universities or hospitals. When their studies involve human participants, they still need to address informed consent, privacy, recruitment, risk, vulnerable populations, and data handling, but often have no institutional IRB or ethics office to guide them.

Today, these researchers may seek an institutional partner, pay a commercial IRB, assemble an ad hoc review process, or proceed without independent review. The problem occurs whenever an unaffiliated researcher proposes human-participant research. The costs can include fees for external review, weeks of delay, substantial administrative work, difficulty publishing or obtaining funding, and most importantly, increased risk to participants if ethical issues are missed.

HP3 begins with Edge Science, including consciousness, anomalous-experience, psi, UAP, and survival-related research. These areas provide a demanding first use case because researchers are often institutionally unaffiliated and studies may involve sensitive experiences, unusual methodologies, privacy concerns, or psychologically vulnerable participants. The underlying problem, however, extends well beyond Edge Science.

2. YOUR WORKFLOW

The trigger is a researcher submitting a study protocol and supporting materials through the HP3 web platform.

1. The researcher creates an account and designates a Research Ethics Coordinator (REC) with current human-participant-protection training.
2. The researcher submits the protocol, consent materials, recruitment language, data-handling information, and relevant supporting documents.
3. HP3 autonomously checks the submission for completeness and identifies missing or conflicting information.
4. The agent classifies the apparent level of risk and determines whether the project is suitable for the HP3 workflow or should be escalated to a conventional IRB or specialist review.
5. For suitable studies, HP3 conducts a structured review covering informed consent, recruitment, risk/benefit, privacy, data handling, vulnerable populations, and study-specific concerns.
6. It generates a review report listing findings, rationale, severity, requested changes, unresolved questions, and suggested improvements.
7. A human REC/reviewer examines the report and can approve, edit, reject, or add findings before they are released.
8. The researcher submits revisions. HP3 compares versions, checks whether findings were addressed, and flags new inconsistencies.
9. A qualified human makes the final determination or escalates the study.

Autonomous work includes document parsing, completeness checks, structured issue detection, source retrieval, version comparison, and draft reporting. Humans retain authority over ambiguous ethical judgments, overrides, escalation, and final determinations.

HP3 currently operates as a dedicated web application. Researchers work in the tools they already use, such as Word, Google Docs, and PDF, and submit those materials through HP3. The Catalyst project would strengthen this into a production-quality workflow rather than a standalone chatbot.

3. TRUST, AUDIT AND GOVERNANCE

Every review will produce a versioned audit trail showing the submitted documents, issues identified, sources consulted, agent recommendations, researcher responses, document revisions, human edits or overrides, and final determination. Users will therefore be able to see both what the agent concluded and why.

Provenance will come from a curated, version-controlled knowledge base of authoritative research-ethics and regulatory guidance. Findings will link back to the rule, guidance, or ethical principle that generated them, and outputs will record the model/workflow version used.

Uncertainty triggers human escalation. Missing critical information, conflicting guidance, higher-risk procedures, vulnerable populations, low-confidence findings, or studies likely to require legally recognized IRB oversight will not receive an autonomous final determination. HP3 is decision support, not a substitute for a legally required IRB.

The human reviewer remains accountable for released findings and final decisions. Validation will measure critical-issue detection, false-positive burden, escalation accuracy, consistency with expert review, and reviewer time saved.

4. TEAM

HP3 is led by Mark Boccuzzi, founder of CEAPAR, co-founder of the Windbridge Institute, and Executive Director of the Windbridge Research Center. His background combines computer science, technology development, research program management, and many years of experience conducting and supporting human-participant research in consciousness and psi.

The project grew directly from that combination of experience. Mark has repeatedly encountered capable independent investigators who want to conduct responsible research but lack the institutional infrastructure available to university researchers.

The project is currently founder-led. There is not yet a formal advisory board. Catalyst funding would allow HP3 to engage compensated research-ethics reviewers and technical specialists and to build a broader reviewer network.

5. WHERE YOU ARE TODAY

HP3 is an early working prototype. A public web platform exists with registration, a document library, submission functionality, and a defined review/revision process. CEAPAR also published a 2024 white paper describing the rationale and early model for AI-assisted ethics review.

The current system has been exercised primarily with sample, hypothetical, and internal cases to develop the workflow. It does not yet have paying customers or a validated user base.

The next step is not simply to add features. We need to test whether HP3 reliably identifies important ethical issues, appropriately escalates cases it should not handle, saves human reviewer time, and produces an audit record that researchers and external stakeholders can trust.

6. ALTERNATIVES AND COMPETITORS

Researchers currently rely on institutional IRBs, commercial/independent IRBs such as Advarra and WCG, informal ethics review, institutional partnerships, or no formal review. Large commercial IRBs provide mature, regulated services, but they are designed primarily for institutional, clinical, and industry research rather than providing lightweight infrastructure for unaffiliated investigators.

AI-assisted IRB pre-review is also emerging. For example, the University of Alabama at Birmingham is piloting an AI IRB document-review tool that helps investigators identify common submission issues before formal review.

HP3 is different in its target user and architecture. It is being designed for researchers who may have no institutional ethics infrastructure at all. It combines multi-step AI review, human oversight, risk-based escalation, iterative researcher/reviewer collaboration, and a persistent audit trail. Edge Science is the initial deployment community and stress-test environment, not the eventual limit of the platform.

7. WHERE THIS GOES

Our longer-term goal is to make credible research-ethics infrastructure available to researchers who currently fall outside institutional systems. HP3 could eventually support independent researchers, community science projects, small nonprofits, research institutes, and organizations that want an AI-assisted ethics pre-review layer.

The platform could also produce portable ethics records that help journals, funders, and collaborators understand what was reviewed, what concerns were raised, what changed, and who approved the final outcome.

HP3 is not currently a commercial product. We want basic access to remain free or very low-cost for independent researchers where possible. Longer-term sustainability could come from institutional subscriptions, private deployments, API/workflow integrations, or paid human-review services.

8. FIT WITH DIGITAL SCIENCE

HP3 sits at the research-integrity and study-planning stage of the research lifecycle, before recruitment and data collection. Its primary users are researchers, with potential downstream value for institutions, publishers, and funders.

The project fits the Catalyst theme because HP3 is genuinely agentic: it performs a sequence of dependent tasks rather than answering a single prompt. It checks a submission, retrieves relevant guidance, evaluates multiple ethical dimensions, requests revisions, compares versions, and prepares a decision record. At consequential points it deliberately transfers control to a human.

We want to work with Digital Science because the challenge is no longer proving that an LLM can comment on an ethics protocol. The challenge is turning that capability into a trustworthy research workflow with provenance, governance, measurable performance, and integration into real research practice. Digital Science’s experience building research infrastructure and trusted research tools would be especially valuable at this transition point.

9. BUDGET

We request £25,000: £9,000 for software engineering and agent orchestration; £4,000 for the curated ethics knowledge and provenance layer; £4,000 for expert reviewer honoraria and reference-standard reviews; £3,000 for validation and red-team testing; £2,500 for AI, hosting, secure storage and backups; £1,500 for privacy, security, accessibility and UX review; and £1,000 for documentation and pilot-user recruitment. This would turn the prototype into a testable, auditable human-AI workflow.