Request for Proposals (RFP)

CAROUSEL invites proposals for a federally aligned research initiative examining how AI-enabled data-acquisition ecosystems may create behavioral-exploitation vectors, peer-to-peer coercion dynamics, and retrocausal security risks. This research will form the foundation for a proposal to entities focused on public safety and national security. The objective is to understand and mitigate emerging threat pathways driven by AI-enabled tools and intelligence. Proposers should demonstrate expertise in AI governance, security studies, behavioral science, insider-threat analysis, temporal-logic modeling, or related fields.

This RFP is a pre-solicitation research call intended to identify collaborators for a future federal or federal-adjacent submission. Selected proposers will be invited to co-develop a joint submission.

AI-enabled platforms increasingly solicit sensitive personal data directly from individuals, offering compensation for access to health, wellness, or behavioral data. There have been observed instances in which these solicitations occur immediately following adverse financial events (e.g., layoffs, job loss, and similar forms of economic destabilization).

This pattern raises several security-related concerns:

A. Algorithmic Vulnerability Detection

AI systems can infer economic stress, employment instability, or behavioral shifts and use these signals to identify individuals most vulnerable to data-solicitation outreach.

B. Human-in-the-Loop Exploitation

Analysts or recruiters who themselves may be operating under economic pressure may escalate outreach, repeatedly target vulnerable individuals, or pressure peers into remaining in compromised roles.

C. Escalation Vectors

Initial low-risk data requests may normalize compliance, creating a pathway toward increasingly sensitive or security-relevant disclosures.

D. Insider-Threat Amplification

Individuals with access to confidential or security-sensitive information may be more susceptible to exploitation during periods of destabilization. These dynamics represent an emerging security risk that has not been systematically studied.

Contractor-heavy ecosystems often place economically vulnerable individuals in competitive or hierarchical roles. When two individuals are simultaneously vulnerable, their incentives may become misaligned with security-related compliance requirements.

This can produce horizontal coercion, in which vulnerable individuals unintentionally exert pressure on one another. Observed behaviors include:

• persuading peers to remain in roles presenting conflicts of interest
• challenging legitimate ethical or legal concerns
• suppressing compliance safeguards
• escalating risks behavior to preserve personal standing
• prioritizing personal survival over organizational obligations

These peer-level interactions constitute a novel insider-threat pathway that traditional security frameworks do not address. CAROUSEL seeks proposals that analyze how these dynamics arise, how they can be detected, and how governance architectures can mitigate them.

CAROUSEL's governance architecture incorporates retrocausal security, defined as:

The study of how anticipated future vulnerabilities, adversarial behaviors, or system states influence present-tense security posture and decision-making.

Retrocausal security applies in this RFP in two ways:

A. Future-State Exploitation

If AI systems or human intermediaries anticipate future economic instability (e.g., layoffs, contract terminations) they may pre-position outreach before destabilization fully manifests.

b. Future-State Harm Propogation

Small, seemingly harmless data requests today may create behavioral patterns that increase susceptibility to more serious breaches in the future. Proposers should articulate how temporal-logic modeling, anticipatory threat frameworks, or future-state constraint analysis can be applied to this domain.

CAROUSEL invites proposal addressing the following research areas:

A. Vulnerability-Indexed Targeting

• Analysis of outreach timing relative to layoffs, off-boarding, or contract termination.
• Identification of demographic or cognitive profiles disproportionately targeted.
• Mapping algorithmic signals used to infer vulnerability
• Algorithmic nudging.

B. Escalation Pathways

• Behavioral-conditioning analysis
• Modeling how benign requests evolve into sensitive asks
• Identification of early-signal indicators
• Identity disruption

C. Inter-Vulnerability Dynamics

• Peer-to-peer coercion modeling
• Contractor-ecosystem risk analysis
• Conflict-of-interest amplification

D. National-Security Implications

• Insider-threat escalation
• Security-relevant data-sharing risks
• Vulnerability-indexed recruitment patterns
• Retrocausal threat modeling

E. Governance and Mitigation Frameworks

• Policy recommendations
• Architectural safeguards
• Early-signal detection systems
• Contractor-ecosystem resilience strategies

Proposers should outline methodological approaches such as mixed-methods analysis, temporal-logic modeling, agent-based simulation, or computational vulnerability-indexing.

Proposers should rely on publicly available datasets, synthetic data, or independently accessible sources. Any proposed data collection must adhere to federal privacy, ethics, and human-subjects standards.

Proposals should include:

• Detailed research plan
• Methodology for data collection and analysis
• Retrocausal security modeling approach
• Threat-model diagrams
• Identification of governance gaps
• Recommendations for federal policy and oversight

Proposers may reference the following foundational works:Madden, Gilman, Levy, Marwick (2017) -- algorithmic vulnerability and digital discrimination

• Siegmund (Naval Postgraduate School, 2025) -- behavioral economics and insider-threat susceptibility
• Sabri et al, USENIX Security (2026) -- economic precarity and peer-level exploitation

A full annotated bibliography is available upon request .

CAROUSEL welcomes submissions from:

• Federally funded research centers (FFRDCs)
• University-affiliated research labs
• Security researchers
• AI governance experts
• Behavioral scientists
• Insider-threat analysts
• Independent researchers with relevant expertise

Proposers should submit the following material to support@carousel.one:

• A 2-4 page concept note
• CVs or bios of principal investigators
• Timeline and budget
• Prior relevant work
• Statement of alignment with federal mission priorities

Formatting Requirements:

• PDF format
• 11-12 pt font
• 1-inch margins
• Section headers matching the Scope of Work
• Maximum 4 pages excluding bios

Notifications will be issued within 14 days of submission. Submitted materials remain the intellectual property of the proposer. CAROUSEL claims no ownership over submitted concepts unless mutually agreed in writing. 

Proposals will be evaluated based on:

• Rigor of methodology
• Novelty of retrocausal security framing
• Feasibility of research plan
• Relevance to security-related concerns
• Alignment with CAROUSEL's governance-layer mission
• Potential for integration into a federal or federal-adjacent program (e.g., DARPA, IARPA).