Position Paper
The DAMREV Position Paper on AI in Compliance Monitoring for Tokenized Instruments
Where rules enforced on-chain end and AI monitoring begins: AML, KYC and transaction monitoring, and who decides
This paper sets out where rules enforced on-chain end and AI monitoring begins, who decides what, and what the law in the group's twelve jurisdictions says about decisions a model makes or shapes. It surveys what the standard setters, supervisors and courts have published, and the enforcement record.
A model may rank, match and explain, but a named person at the licensed entity decides whether a customer is refused, restricted or reported.
Version 1.0Date 16 November 2026Pages 79Sections 16Sources 60

At a glance
Who decides?
AI may assist screening, risk assessment and monitoring. It never changes a holder's transfer permissions on its own, and a named person at the licensed entity decides whether a customer is refused, restricted or reported.
- The ruleApproved rules, no model inside.
- The modelRanks, matches and surfaces activity.
- The personDecides, records and answers.
What is inside
What the paper covers
- The boundary in three layersAML work for a tokenized instrument falls into the encoded rule, screening and monitoring. AI is absent from the first, assists the second and does most of the work of the third.
- Monitoring on a public ledgerA public ledger shows every transfer of the instrument. It does not show who owns an address: analytics turns a wallet into a probable owner, and a licensed person decides whether probable is enough to act on.
- Decisions about peopleThree bodies of law reach a decision about a customer when a model makes or shapes it: the EU AML Regulation, data protection law and the AI Act. In nine of the twelve jurisdictions a data protection statute gives individuals rights concerning decisions based solely on automated processing.
- The human in the loopA person at every discretionary decision is necessary. It stops at mandatory restrictions, and it is not sufficient on its own: a review that confirms a model's output at a rate no person could sustain is automation with a signature on it.
- The enforcement recordIn the most cited cases the failures were of coverage, capacity and configuration. None turned on whether a firm used AI.
- Monitoring data and the ledgerMonitoring produces the most sensitive records a programme holds, and none of them belongs on a ledger. Under the group's design, scores, alerts, case notes and reports stay off-chain with the licensed entity, and any restriction visible on the ledger is assessed for what it reveals.
- Version
- 1.0
- Date
- 16 November 2026
- Pages
- 79
- Sections
- 16
- Sources
- 60
The positions
The seven positions
- Code executes; models inform
- No model runs inside the transfer engine
- Discretionary decisions about a customer are human; mandatory restrictions are executed by approved rule and reviewed
- The licensed entity answers for the model's outcome, whoever built it
- Coverage comes before sophistication
- Monitoring records stay off the ledger
- DAMREV describes a design, not its effectiveness or a product's compliance
The paper's key figure
The position on one page
Figure 15. The whole position on one page. A design requirement, not a description of operations in place.
Step 1
The rule
Sentinel executes approved rules at every transfer, with no model inside.
Step 2
The model
Screening and monitoring models rank, match and surface activity for review.
Step 3
The person
A named person at the licensed entity decides, records and answers for every discretionary outcome.
Step 4
Mandatory restrictions
Execute at once under a rule approved in advance; reviewed after.
Step 5
Model risk
Inventory, pre-use tests, triggers, both error measures, fairness, change control.
Step 6
Data
CDD data limits honoured; records off the ledger; restrictions assessed for inference.
Contents
How is the paper laid out?
16 sections. The opening of Section 01 is open to read here. The full paper comes by email.
- Executive summaryp. 9
- Seven positionsp. 13
- 01The 11 findingsOpenp. 17
- 02Two kinds of controlp. 20
- 03The boundary in three layersp. 22
- 04AI at onboarding: identity, liveness, and the deepfakep. 25
- 05Monitoring on a public ledgerp. 27
- 06What the standard setters and supervisors sayp. 30
- 07Decisions about peoplep. 33
- 08The human in the loopp. 36
- 09The enforcement recordp. 39
- 10Model risk and validationp. 42
- 11Monitoring data and the ledgerp. 45
- 12The 12 jurisdictionsp. 47
- 13The DAMREV monitoring architecturep. 50
- 14Claim disciplinep. 53
- 15Questions for counselp. 55
- 16Sourcesp. 58
- The position on one page & Appendicesp. 57
Read the opening of Section 01
This paper surveys the use of artificial intelligence in anti-money laundering, know your customer and transaction monitoring for tokenized instruments from primary sources: what the standard setters, supervisors and courts have published, the binding law in each of the group's twelve jurisdictions, and the enforcement record. It reaches seven positions.
1. Code executes; models inform. Under the group's design, the transfer engine applies approved eligibility, transfer and sanctions rules deterministically. AI may assist screening, risk assessment and monitoring. It never changes a holder's transfer permissions on its own.
The rest of the paper is sent by email. Request the paper
How to read it
How do you get the full paper?
The same four steps as every DAMREV paper. The paper is sent by email, so each copy goes to a named reader.
Step 1 · You are here
Read the preview
The paper's page carries the summary, the contents and who the paper is for.
Step 2
Request the paper
Give your name, work email and organisation, and consent to receive it.
Step 3
Open the link
A link arrives by email. It expires 24 hours after it is sent.
Step 4
Read the paper
The link opens the paper in a protected reader on damrev.com, for reading only.
Request the paper
Where do we send the paper?
The form is the same one used for every paper. Further contact is optional, and the paper is sent either way.
Request the AI in Compliance Monitoring Position Paper

What you receive
- One email from DAMREV with a personal link to this paper.
- The link opens the paper in a protected reader on damrev.com. It is for reading only and cannot be downloaded or printed.
- The link expires 24 hours after it is sent. You can request a new one at any time.
- The email carries the link only. Anything further needs your separate consent.
Common questions
What else do people ask?
Short answers to the questions people ask most about this topic.
Why is the paper sent by email?
So that each copy goes to a named reader. The request works the same way for every DAMREV paper.
What if my link has expired?
Request the paper again from this page, and a new link is sent to the address you give.
Do I have to agree to further contact?
No. Further contact is optional, and the paper is sent either way.
Does an AI model run inside the transfer engine?
No. Under the group's design, the transfer engine applies approved rules deterministically and does not run an AI model. This is the group's design choice, not a claim that the law forbids it; a model-derived restriction reaches eligibility only through an authorised control process, with an approver, a reason, a version and an audit trail.
Who decides whether a customer is refused, restricted or reported?
Onboarding refusals, exits, changes in due diligence and restrictions based on suspicion are decided by a named person at the licensed entity. A confirmed sanctions match or a lawful order is executed at once by a rule a person approved in advance, and reviewed after. Reporting a suspicion is the reporting officer's decision, under its own law.
Can blockchain analytics show who owns a wallet?
Blockchain analytics infers ownership; it does not observe it. Attribution is an inference with a stated degree of confidence, and a licensed person decides whether probable is enough to act on.
Are monitoring records written to the ledger?
No. Under the group's design, risk scores, alerts, case notes and suspicious activity reports are never written on-chain, in clear or in a form that could be linked to a wallet. They are processed only for compliance, regulatory and legal purposes, never for commercial ones.
Does the paper show that the controls work?
No. The controls in this paper are the group's design requirements. They are not evidence that any implementation satisfies the law or works, which only instrument-specific legal review and operating evidence can show.
Is the paper advice or an offer?
No. The paper is provided for information only. It is not an offer of any token or financial instrument, and it is not legal, tax or investment advice.
Where next
Where should you go next?
The papers and pages below carry the detail this paper leaves to them. Each paper is requested from its own page.
- Compliance by Design Position PaperWriting transfer rules, eligibility conditions, and intervention powers into code: what code can enforce, and what only law and licence decide.
- Screening & OnboardingHow investors are verified by a licensed provider before holding anything and checked again at every transfer, with anti-money laundering, sanctions, and travel-rule obligations met.
- SentinelSentinel decides whether a transfer may proceed, reading eligibility from Veritas, enforcing the asset's rules on-chain, and blocking any transfer that fails.
- Governing AI Position PaperWhat the law already requires when a regulated firm uses artificial intelligence, what the new instruments add, and the practice the group holds itself to.
References
- 014.Regulation (EU) 2024/1624 (AML Regulation), Article 76. https://eur-lex.europa.eu/eli/reg/2024/1624/oj
- 006.FATF, Opportunities and Challenges of New Technologies for AML/CFT, July 2021. https://www.fatf-gafi.org/en/publications/Digitaltransformation/Opportunities-challenges-new-technologies-for-aml-cft.html
- 009.Wolfsberg Group, Principles for Using Artificial Intelligence and Machine Learning in Financial Crime Compliance, 1 December 2022. https://wolfsberg-group.org/news/34
Last updated:






