Free private markets document template
AI Policy Template for LPs
Create a private markets AI policy covering permitted uses, confidential data, human review, vendors, model risk, records and incident response.
01
Purpose
This policy governs responsible artificial intelligence use by Pension Investment Office effective October 1, 2027.
02
Scope
Covered users are trustees, employees, consultants, contractors and service providers. Covered activities include manager research, document review, data extraction, monitoring, operations, reporting and communications.
03
Core principles
Use AI to augment accountable professionals. Apply confidentiality, accuracy, transparency, proportionality, security and human judgment to every material use.
04
Permitted uses
Permitted uses include summarization, draft generation, document classification, controlled extraction, coding support and internal search, subject to approved tools and appropriate review.
05
Prohibited uses
autonomous investment approval, unsupervised valuation, fabricated citations, uploading restricted data to unapproved tools and impersonation
06
Confidential data
Use only approved enterprise tools with contractual privacy, tenant isolation and no training on institutional data
07
Human oversight
A qualified employee verifies material facts, calculations, sources and recommendations before use
08
Vendor diligence
Security, privacy, model documentation, data location, subcontractors, resilience and exit rights
09
Model risk
Test accuracy, hallucination, bias, drift, explainability and reproducibility before production use
10
Security
SSO, role-based access, logging, retention limits, data-loss prevention and incident monitoring
11
Records & disclosure
Retain material prompts, outputs, sources, reviewer and final disposition for decision-relevant uses
12
Incident response
Stop affected use, preserve evidence, notify Security and Legal, assess impact and remediate
13
Governance
AI Governance Committee with Investments, Operations, Legal, Risk and Information Security
14
Training
Training is Required before access and refreshed annually. Users must understand approved tools, prohibited data, verification and incident reporting.