The Financial Action Task Force sets the global AML/CFT standard that national regimes implement, and it has never written an AI rule: its position is that the technology-neutral Recommendations already apply, with Recommendation 15 (new technologies) requiring institutions to assess the risks of any new or developing technology before launch, Recommendation 1 requiring proportionate, risk-based measures, Recommendation 10 governing customer due diligence and Recommendation 20 the reporting of suspicions. Its AI-relevant documents are the March 2020 Guidance on Digital Identity, the two July 2021 digital-transformation reports (new technologies for AML/CFT; data pooling and collaborative analytics) and the December 2025 horizon scan on AI and deepfakes, which warns that generative AI, AI agents and deepfakes are already used to defeat customer due diligence and transaction monitoring. Nothing binds a bank directly; everything reaches it through FinCEN, the EU AML framework, the UK regulations and other national implementations.
| Full name | Financial Action Task Force (FATF) |
| Role | Global AML/CFT standard-setter |
| Force on banks | Non-binding standards |
| Applies to | FATF member jurisdictions, which implement the Recommendations in national law and are assessed against them; banks, virtual asset service providers and other reporting entities through that national law (FinCEN and the Bank Secrecy Act in the US, the EU AML framework, the UK Money Laundering Regulations) |
| Key document | Horizon Scan: Artificial Intelligence and Deepfakes (December 2025), read with Recommendation 15 of the FATF Recommendations (consolidated October 2025) |
| Latest move | Dec 22, 2025: horizon scan on AI and deepfakes published after approval at the 22–24 October 2025 Plenary; the February and June 2026 Plenary outcomes record no further AI item |
| Documents tracked | 5 · all documents → |
The FATF's output on technology comes in two kinds. The first is the Standards themselves — the 40 Recommendations adopted in February 2012 and consolidated most recently in October 2025 — which are deliberately technology-neutral: Recommendation 15 asks countries and financial institutions to identify and assess the risks of new products, business practices and 'new or developing technologies' before launch, and the February 2025 revision of Recommendation 1 and the interpretive notes to Recommendations 10 and 15 pushed proportionality and simplified measures in lower-risk areas. The second kind is guidance and reports that say how the Standards apply: the 2020 digital identity guidance built an assurance-level test for digital onboarding under Recommendation 10, and the July 2021 reports endorsed machine learning, advanced analytics and privacy-preserving collaborative analytics for AML/CFT provided they are explainable, human-overseen, privacy-respecting and cyber-secure.
Since 2025 the FATF has treated AI as a threat as well as a tool. The horizon scan approved at the October 2025 Plenary and published on 22 December 2025 describes deepfakes defeating biometric and video KYC, AI agents running automated laundering pipelines and synthetic transaction patterns designed to look legitimate, and concludes that these tools 'challenge existing AML/CFT/CPF frameworks, particularly in areas like customer due diligence, identity verification, and transaction monitoring'. It still issues no new requirement: the response is Recommendation 15 risk assessment, stronger verification and responsible use of AI in detection. The February and June 2026 Plenary outcomes record no further AI item, so the horizon scan is the FATF's current word.
What has the FATF actually published on AI?
| Date | Document | Status |
|---|---|---|
| Dec 22, 2025 | FATF horizon scan on AI and deepfakes (2025) — Horizon Scan: Artificial Intelligence and Deepfakes — Impacts on Money Laundering, Terrorist Financing and Proliferation Financing | Final |
| Jul 1, 2021 | FATF new technologies for AML/CFT report (2021) — Opportunities and Challenges of New Technologies for AML/CFT | Final |
| Jul 1, 2021 | FATF data pooling and collaborative analytics stocktake (2021) — Stocktake on Data Pooling, Collaborative Analytics and Data Protection | Final |
| Mar 6, 2020 | FATF Guidance on Digital Identity (2020) — Guidance on Digital Identity | Final |
| Feb 15, 2012 | FATF Recommendations — International Standards on Combating Money Laundering and the Financing of Terrorism & Proliferation: The FATF Recommendations (consolidated text updated October 2025) | In force |
| Date | Type | Document / event |
|---|---|---|
| Dec 22, 2025 | Report | FATF horizon scan on AI and deepfakes (2025) — Horizon Scan: Artificial Intelligence and Deepfakes — Impacts on Money Laundering, Terrorist Financing and Proliferation Financing. The FATF's horizon scan on artificial intelligence and deepfakes, approved at the 22–24 October 2025 Plenary and published on 22 December 2025, is the standard-setter's first document on generative AI, AI agents and deepfakes as money-laundering, terrorist-financing and proliferation-financing risks. source ↗ |
| Oct 24, 2025 | Milestone | October 2025 Plenary approves the AI and deepfakes horizon scan. The 22–24 October 2025 Plenary 'approved a new Horizon Scan, to notify public and private sectors around the world about current and potential future illicit finance risks presented by artificial intelligence (AI) and deepfakes', warning that criminals 'can exploit generative AI, AI agents and other new technologies'; the scan was published on 22 December 2025. source ↗ |
| Feb 2025 | Milestone | Recommendation 1 and the interpretive notes to Recommendations 10 and 15 revised for proportionality. The February 2025 Plenary revised Recommendation 1, INR.1, INR.10 and INR.15 'to increase focus on proportionality of measures and require countries to allow and encourage simplified measures in lower risk areas', the change most relevant to risk-based, technology-assisted due diligence. source ↗ |
| Jul 1, 2021 | Report | FATF new technologies for AML/CFT report (2021) — Opportunities and Challenges of New Technologies for AML/CFT. The FATF's report 'Opportunities and Challenges of New Technologies for AML/CFT', finalised at the June 2021 Plenary and published on 1 July 2021, is the standard-setter's position that technology — digital identity, transaction monitoring and analysis solutions including machine learning, and collaborative analytics — can make AML/CFT measures faster, cheaper and more accurate when used responsibly and on a risk-based basis. source ↗ |
| Jul 1, 2021 | Report | FATF data pooling and collaborative analytics stocktake (2021) — Stocktake on Data Pooling, Collaborative Analytics and Data Protection. The FATF's stocktake, finalised at the June 2021 Plenary and published on 1 July 2021, examines commercially available and emerging technologies that let financial institutions run advanced AML/CFT analytics on large structured and unstructured datasets and, more ambitiously, analyse data collaboratively across institutions while respecting data-protection and privacy law. source ↗ |
| Jun 25, 2021 | Milestone | June 2021 Plenary finalises the two digital-transformation reports. The fourth Plenary under the German Presidency, held 21–25 June 2021, completed the new-technologies report and the data-pooling stocktake, both published on 1 July 2021, and 'discussed progress on the use of advanced analytics and machine learning in detecting suspicious activities of money laundering and terrorist financing'. |
| Mar 6, 2020 | Guidance | FATF Guidance on Digital Identity (2020) — Guidance on Digital Identity. The FATF's Guidance on Digital Identity, published on 6 March 2020, explains how regulated entities can use digital ID systems to meet the customer identification and verification requirement of Recommendation 10(a) and to support ongoing due diligence under Recommendation 10(d). source ↗ |
| Feb 15, 2012 | Framework | FATF Recommendations — International Standards on Combating Money Laundering and the Financing of Terrorism & Proliferation: The FATF Recommendations (consolidated text updated October 2025). The FATF Recommendations are the global AML/CFT standard that national regimes (the US Bank Secrecy Act, the EU anti-money-laundering framework, the UK Money Laundering Regulations) implement; adopted by the FATF Plenary on 15 February 2012, the consolidated text is marked 'Updated October 2025'. source ↗ |
- Whether the FATF follows the December 2025 horizon scan with guidance or a revision to the interpretive note to Recommendation 15 on AI, deepfakes and agents; the February and June 2026 Plenary outcomes record no such project
- Any refresh of the 2020 Guidance on Digital Identity for synthetic-media and injection attacks on biometric and video verification, which the horizon scan names as the main vulnerability
- Outputs of the UK Presidency (2026–2028) on technology and AML/CFT, and whether the FATF's digital-transformation workstream resumes after the 2021 reports
- How national implementers — FinCEN's AML/CFT programme rule, the EU AML Authority, the UK regulations, the HKMA's AI-in-AML work — cite the FATF texts when they set expectations for AI monitoring and SAR tools
Does the FATF regulate AI in anti-money laundering?
No. The FATF sets non-binding global standards and has issued no AI-specific standard. Its position is that the technology-neutral Recommendations apply: Recommendation 15 requires a risk assessment of any new or developing technology before launch, and Recommendations 1, 10 and 20 govern proportionality, customer due diligence and suspicious-transaction reporting. Its AI-relevant texts are the 2020 digital identity guidance, the 2021 new-technologies and data-pooling reports and the December 2025 horizon scan on AI and deepfakes.
What does the FATF say about using machine learning for transaction monitoring?
The July 2021 report 'Opportunities and Challenges of New Technologies for AML/CFT' says machine learning, advanced analytics and collaborative analytics can make AML/CFT measures faster, more accurate and more inclusive when used responsibly and on a risk-based basis, and its suggested actions ask for explainability and transparency, human oversight, privacy and data protection, strong cybersecurity and alignment with standards. The 2021 data-pooling stocktake adds that cross-institution analytics are compatible with data-protection law when privacy-enhancing technologies are used.
What is the FATF horizon scan on AI and deepfakes?
A report approved at the October 2025 Plenary and published on 22 December 2025 that describes how deepfakes defeat customer due diligence and biometric verification, how AI agents and predictive models could automate layering and mimic legitimate transactions, and how institutions and authorities are detecting it. It creates no new obligation and anchors the response in Recommendation 15.
Follow every move these regulators make
when one of these regulators moves, the next morning's brief says so · six sourced stories · 7 am ET · free
plus every tracker, bank and agent page update, the morning after · leave any morning