Effective Date: Sep 3, 2025 | Last Updated: Dec 18, 2026
Jade Woods Venture Capital partners with founders building the neural infrastructure of the 21st century. We believe artificial intelligence can be a powerful multiplier of human capability, scientific progress, enterprise productivity, and creative expression. We also recognize that AI systems can create material risks to privacy, security, safety, fairness, accountability, labor, democratic integrity, and human agency if they are designed, deployed, or funded without appropriate governance.
This AI Ethics Charter sets out the principles, expectations, and internal governance commitments that guide Jade Woods’ investment process, portfolio engagement, and internal use of AI. It is intended to support responsible innovation, not to slow ambitious technical progress. Our conviction is straightforward: the most durable AI companies will combine technical superiority with defensible governance.
1. Scope and Status
This Charter applies to Jade Woods’ evaluation of AI-related investments, engagement with portfolio companies, internal operational use of AI tools, public communications about AI, and strategic partnerships involving AI systems or AI infrastructure.
The Charter is a governance and values document. It does not create a contractual warranty, fiduciary duty, investment commitment, public representation about any specific portfolio company, or private right of action. Specific obligations may be created only through definitive agreements, fund documents, side letters, investment documents, vendor contracts, employment policies, or other written instruments.
2. Definitions
For purposes of this Charter:
“AI system” means a machine-based system that can, for explicit or implicit objectives, infer from inputs how to generate outputs such as predictions, content, recommendations, decisions, classifications, or actions that may influence physical or digital environments.
“Generative AI” means AI capable of producing text, code, images, audio, video, synthetic data, 3D assets, designs, simulations, or other content.
“Agentic AI” means AI capable of planning, executing, or coordinating multi-step tasks with varying levels of autonomy, tool use, memory, delegation, or environmental interaction.
“High-impact AI” means AI that materially affects rights, safety, access to essential services, financial outcomes, employment, education, healthcare, housing, law enforcement, critical infrastructure, democratic processes, or other consequential domains.
“Responsible AI” means AI designed, developed, deployed, monitored, and governed in a way that is lawful, safe, secure, privacy-protective, transparent, accountable, robust, and aligned with human dignity and legitimate user expectations.
3. Core Principles
A. Human Agency and Dignity
AI systems should augment human capability and preserve meaningful human agency. AI should not be designed to manipulate people through deception, exploit vulnerabilities, remove appropriate human oversight from consequential decisions, or undermine human dignity. Founders should be clear about when humans remain accountable for decisions and where human review, override, escalation, or appeal is required.
B. Privacy and Data Governance
AI companies should treat privacy and data governance as architectural requirements, not compliance afterthoughts. We expect appropriate data minimization, lawful collection, informed use, secure storage, retention controls, access management, provenance tracking, consent where required, and reasonable restrictions on secondary use. Sensitive, regulated, proprietary, or confidential data should receive heightened safeguards.
C. Safety, Security, and Robustness
AI systems should be designed and evaluated for foreseeable misuse, failure modes, prompt injection, data leakage, model extraction, cyber abuse, adversarial manipulation, hallucination, unsafe autonomy, overreliance, and operational resilience. The depth of testing and monitoring should scale with the system’s capability, deployment context, user base, and potential impact.
D. Fairness and Non-Discrimination
AI systems should be assessed for harmful bias, disparate impact, exclusion, and unjustified performance gaps across relevant groups and contexts. Fairness analysis should be tied to the actual use case and affected populations. Companies should document known limitations and implement mitigation strategies where the system can materially affect people’s opportunities, access, safety, or rights.
E. Transparency, Explainability, and User Disclosure
Users and affected parties should receive meaningful information about AI system capabilities, limitations, data inputs, human oversight, and appropriate use. AI-generated or AI-mediated content should be disclosed where disclosure is required by law or where nondisclosure would be deceptive or materially misleading. Explainability should be proportionate to the consequences of the system and the needs of the audience.
F. Accountability and Traceability
Responsible AI requires clear owners, escalation paths, audit trails, documentation, incident handling, and governance records. Companies should be able to explain who is accountable for system design, dataset selection, model behavior, deployment decisions, monitoring, remediation, and user communications. Traceability should support technical debugging, regulatory response, contractual compliance, and user trust.
G. Scientific Integrity, IP, and Data Provenance
AI companies should respect intellectual property, trade secrets, data rights, licensing restrictions, confidentiality obligations, and research integrity. Claims about model performance, autonomy, safety, accuracy, cost, latency, compute efficiency, and benchmarks should be supportable. Training and evaluation datasets should be documented to a level appropriate for the business model, technical risk, and legal environment.
H. Environmental Responsibility and Compute Efficiency
Compute is a strategic asset and an environmental consideration. AI infrastructure companies should pursue efficient architectures, responsible resource allocation, measurement of energy-intensive workloads where feasible, and product choices that reduce unnecessary compute waste without compromising safety or reliability.
I. Lawfulness and Global Interoperability
AI companies should understand the regulatory regimes relevant to their product, customers, deployment geographies, and data flows. This may include privacy, cybersecurity, consumer protection, employment, financial services, healthcare, export control, sanctions, copyright, product liability, competition, securities, and AI-specific laws. Responsible AI governance should be globally aware and adaptable as legal standards evolve.
4. Investment Due Diligence Framework
Jade Woods applies a risk-proportionate AI diligence approach. The depth of diligence depends on the company’s stage, sector, autonomy level, model capability, data sensitivity, user base, and potential impact.
A. Baseline Questions
For AI-related opportunities, investment teams should consider:
What problem does the AI system solve, and who may be affected by its outputs or actions?
What data is collected, licensed, generated, inferred, retained, or shared?
What model architecture, foundation model, dataset, vendor, infrastructure, and deployment environment are used?
What are the system’s known limitations, failure modes, uncertainty boundaries, and unsafe-use scenarios?
How are security, privacy, access control, monitoring, logging, and incident response handled?
Does the system materially affect rights, safety, essential services, financial outcomes, employment, education, healthcare, housing, law enforcement, critical infrastructure, or democratic processes?
What claims are being made to customers, regulators, investors, or the public, and are those claims supportable?
What governance, documentation, testing, red teaming, evaluation, and human oversight practices are in place?
What legal regimes apply today, and what credible regulatory changes could affect the company within the next 12 to 36 months?
B. Risk Tiers
For internal analysis, Jade Woods may classify AI opportunities into four indicative categories:
Low-risk AI: internal productivity tools, developer tooling, infrastructure, or applications with limited effect on individuals and low foreseeable harm.
Moderate-risk AI: systems that influence business workflows, user recommendations, content generation, enterprise decisions, or customer operations but retain clear human oversight and limited consequential impact.
High-impact AI: systems affecting regulated, safety-critical, rights-impacting, or consequential domains, including healthcare, finance, employment, education, insurance, law, security, critical infrastructure, robotics, autonomous agents, or large-scale content distribution.
Prohibited or presumptively excluded AI: systems primarily designed for unlawful activity, deception, social scoring, exploitative manipulation, illegal surveillance, non-consensual biometric identification, deepfake abuse, discrimination, unauthorized cyber operations, fraud, market manipulation, or other uses that Jade Woods determines to be inconsistent with this Charter.
This risk-tier framework is an internal guide and does not replace legal analysis. A company’s risk tier may change as the product, users, geography, capabilities, or regulatory environment changes.
C. Red Flags Requiring Escalation
The investment team should escalate opportunities where there are indicators of:
Unclear data rights, questionable scraping practices, insufficient licensing, or reliance on sensitive data without a defensible basis.
High-impact decisions without human oversight, appeal rights, or meaningful explanation.
Material security weaknesses, lack of incident response, poor access controls, or exposure to prompt injection, model extraction, data leakage, or cyber misuse.
Claims of accuracy, autonomy, safety, or compliance that are unsupported by testing or documentation.
Business models dependent on deceptive design, dark patterns, covert surveillance, behavioral manipulation, or regulatory arbitrage.
Use cases involving minors, vulnerable populations, biometric data, health data, financial data, employment decisions, law enforcement, military uses, or critical infrastructure without adequate safeguards.
Founder unwillingness to address responsible AI risks or to adopt basic governance practices as the company scales.
5. Portfolio Engagement Expectations
Jade Woods aims to support portfolio companies in building responsible AI practices appropriate to their stage and risk profile. We may encourage, assist with, or request progress on:
Data maps, model cards, system cards, evaluation reports, architecture documentation, vendor registers, and security documentation.
Privacy-by-design practices, data minimization, access controls, retention schedules, consent management, and data processing agreements.
Safety evaluations, red teaming, adversarial testing, misuse testing, bias testing, human factors testing, and post-deployment monitoring.
User disclosures, acceptable use policies, terms of service, abuse reporting, escalation procedures, and customer-facing AI documentation.
Incident response plans for data breaches, model failures, harmful outputs, security vulnerabilities, misuse, and regulatory inquiries.
Responsible go-to-market practices, including supportable claims, honest benchmarking, transparent limitations, and appropriate customer education.
Regulatory readiness for AI-specific, privacy, cybersecurity, copyright, export control, sanctions, sectoral, and consumer protection obligations.
Governance ownership, board-level or leadership-level reporting, employee training, and periodic review of high-impact AI systems.
6. Internal Use of AI at Jade Woods
Jade Woods may use AI tools for research, productivity, drafting, translation, summarization, diligence support, market mapping, portfolio support, and operational workflows. Internal AI use should follow these rules:
Confidential founder materials, investor information, non-public portfolio information, regulated data, trade secrets, and sensitive Personal Information should not be entered into public AI tools unless the tool has been approved for that data class and appropriate safeguards are in place.
AI-generated summaries, diligence notes, research, and recommendations should be reviewed by a qualified human before being used in investment decisions, external communications, legal documents, or board materials.
AI tools should not be treated as authoritative sources for legal, tax, regulatory, technical security, medical, or financial conclusions without independent verification.
Employees and contractors should preserve records needed for compliance, diligence, conflicts analysis, and auditability when AI materially assists a workflow.
AI outputs should be checked for hallucinations, bias, outdated information, licensing issues, confidential information leakage, and unsupported claims.
7. Governance and Accountability
Jade Woods should designate internal responsibility for AI governance. Depending on firm size and operations, this may be handled by a responsible partner, operating lead, legal/compliance lead, or AI ethics committee. The responsible function should:
Maintain this Charter and review it at least annually or when material legal, technical, or market developments require updates.
Support investment teams in escalating high-impact AI opportunities and red flags.
Maintain internal guidance on approved AI tools, prohibited data inputs, vendor review, and responsible AI workflows.
Coordinate with counsel, technical advisors, security advisors, and domain experts where specialized review is needed.
Encourage training for investment professionals and operating teams on AI risk, privacy, security, model evaluation, and relevant regulation.
Document material exceptions, escalations, and remediation steps in appropriate internal records.
8. Incident Reporting and Remediation
Jade Woods expects material AI incidents to be addressed promptly and proportionately. A material AI incident may include a safety failure, privacy breach, model misuse, harmful bias event, unlawful output, data leakage, security vulnerability, regulatory notice, customer-impacting hallucination, autonomous action failure, or other event that materially affects users, customers, third parties, portfolio value, or legal compliance.
Where Jade Woods becomes aware of a material AI incident involving internal operations or a portfolio company, the appropriate response may include investigation, containment, technical remediation, user notification, customer communication, regulatory consultation, board escalation, vendor review, external expert review, suspension of features, product changes, policy updates, or other corrective actions.
9. Prohibited Uses and Investment Exclusions
Jade Woods will not knowingly support, fund, or deploy AI systems where the primary intended use is:
Unlawful surveillance, unlawful biometric identification, or violation of privacy rights.
Social scoring or rights-impacting profiling that is unlawful, discriminatory, or incompatible with democratic values.
Fraud, phishing, impersonation, financial manipulation, market manipulation, deceptive synthetic media, or abuse of trust at scale.
Generation, distribution, or facilitation of child sexual abuse material, non-consensual intimate imagery, or sexual exploitation.
Malicious cyber activity, unauthorized access, credential theft, malware deployment, vulnerability exploitation, or evasion of security controls.
Unlawful discrimination or denial of rights, services, or opportunities based on protected characteristics.
Autonomous physical harm or weaponization without lawful authority, meaningful human oversight, and rigorous safeguards.
Material violation of sanctions, export controls, human rights law, or other applicable legal restrictions.
This exclusion list is not exhaustive. Jade Woods may decline or disengage from opportunities that, in its judgment, are inconsistent with this Charter or create unacceptable legal, ethical, security, reputational, or societal risk.
10. Practical Implementation Checklist
For Pre-Investment Review
Identify the AI use case, affected stakeholders, autonomy level, and risk tier.
Review data provenance, data rights, privacy posture, and retention practices.
Assess model performance claims, evaluation quality, benchmark integrity, and failure modes.
Review security posture, misuse controls, abuse reporting, and incident response readiness.
Identify applicable AI, privacy, cybersecurity, IP, export, sanctions, and sectoral regulations.
Escalate red flags to the appropriate partner, legal/compliance advisor, or technical expert.
For Portfolio Support
Create a stage-appropriate responsible AI roadmap.
Adopt user-facing terms, privacy notices, acceptable use rules, and AI disclosures where needed.
Implement evaluation, monitoring, red teaming, and incident response processes for high-impact systems.
Prepare documentation for enterprise customers, regulators, auditors, and strategic partners.
Review claims in sales, marketing, fundraising, and public communications for accuracy and supportability.
For Internal Jade Woods Operations
Maintain an approved AI tools list and prohibited input guidance.
Train employees and contractors on confidentiality, privacy, AI hallucination risk, and human review requirements.
Document material AI use in diligence workflows where appropriate.
Review the Charter periodically against legal developments, industry standards, and portfolio learnings.
11. Contact
Questions about this AI Ethics Charter may be directed to:
Jade Woods Venture Capital
Attention: Responsible AI / Ethics
Email: ethics@jadewoods.vc
No Third-Party Rights
This AI Ethics Charter expresses Jade Woods’ responsible AI principles and internal operating expectations. It is not intended to create contractual rights, fiduciary duties, warranties, or obligations to any third party, except where expressly incorporated into a separate written agreement signed by Jade Woods.
Relationship to Other Policies
This Charter should be read together with the Jade Woods Privacy Policy, Terms of Service, applicable confidentiality agreements, investor documents, employment policies, security policies, and portfolio-specific agreements. If a binding agreement imposes stricter obligations, the binding agreement controls.