AI Risk Wire

Risks

Fraud and impersonation

AI is used to pose as a person, brand or institution to take money or access.

NIST AI 600-1 · Information IntegrityFCA Mills Review (2026)EU AI Act Art. 50

21incidents (primary) · 11 more as secondary
0read for controls · 21 not yet
1fixes reported
0repeats after a fix

Incidents per day

0 on 1 Nov 2026; average of the 7 days before: 0.4.

4 Oct 2026: 05 Oct 2026: 06 Oct 2026: 07 Oct 2026: 08 Oct 2026: 19 Oct 2026: 010 Oct 2026: 011 Oct 2026: 112 Oct 2026: 113 Oct 2026: 014 Oct 2026: 015 Oct 2026: 016 Oct 2026: 017 Oct 2026: 018 Oct 2026: 119 Oct 2026: 020 Oct 2026: 121 Oct 2026: 022 Oct 2026: 023 Oct 2026: 024 Oct 2026: 025 Oct 2026: 026 Oct 2026: 027 Oct 2026: 028 Oct 2026: 029 Oct 2026: 130 Oct 2026: 231 Oct 2026: 01 Nov 2026: 02 Nov 2026: 0 6 Oct9 Oct12 Oct15 Oct18 Oct21 Oct24 Oct27 Oct30 Oct2 Novto date

Developers

Perplexity · 1

Where the chain broke

Part of the systemIncidents where the control was not effective or did not existIncidents where the control was effectiveIncidents where sources don't mention the controlIncidents not yet checked for controls
L1 Model
Model provider
21
L2 Harness
Agent framework vendor
21
L3 Application
Deployer
21
L4 Infrastructure / target
Organisation hit
21

Controls for this risk, by layer

ControlIncidents where the control was not effective or did not existIncidents where the control was effectiveIncidents where sources don't mention the controlMost common reason it was not effective
Content safety classifiers
L1 Model
Abuse monitoring and account enforcement
L1 Model
Marking machine-made output
L1 Model
People told they are dealing with AI
L3 Application
Out-of-band verification of payment and identity requests
L4 Infrastructure / target
Phishing and fraud filters
L4 Infrastructure / target

6 of 6 controls for this risk not yet recorded as effective.

Fixes reported and what followed

IncidentControl added by the fixIncidents after the fix
LR-2026-0723 2026-05-19
Fix reported
Not tied to a control

Incidents

EffectiveNot effectiveNot effective (bypassed)Did not existNot stated in sourcesNot yet checkedAttacker's ownOrder: model · harness · application · target

DateIncidentRiskWhere it failedControls mentioned in sources
2026-10-30Fake Polymarket platforms use LLM chat agents with provider-switching to defraud victims Fraud and impersonation Not yet checked
2026-10-29North Korean IT worker scheme uses FFTDC 'Super Intelligence' AI tools, deepfakes and voice manipulation to defeat U.S. hiring controls Fraud and impersonation Not yet checked
2026-10-20AI-generated 'Xiaohai' teen persona used to sell products to older women via fabricated sob stories Fraud and impersonation Not yet checked
2026-10-18Deepfake video of President Marcos used to defraud Tarlac dentist of P93 million Fraud and impersonation Not yet checked
2026-10-12Fideuram (Intesa Sanpaolo private banking) deepfake voice/video executive impersonation fraud Fraud and impersonation Not yet checked
2026-10-11AI-morphed obscene image used to blackmail 19-year-old in Mangaluru Fraud and impersonation Not yet checked
2026-10-08Deepfake investment-scam ads impersonating Nigerian public figures on Facebook Fraud and impersonation Not yet checked
2026-09-14Fake YouTube stream uses AI Trump video to promote crypto scam Fraud and impersonation Not yet checked
2026-08-03Deepfake celebrity likenesses used in AI scam pushing fake supplements Fraud and impersonation Not yet checked
2026-07-31North Korean IT workers use real-time deepfakes to beat hiring checks Fraud and impersonation Not yet checked
2026-07-17Sky Gold subsidiary suffers loss after deepfake scam and unauthorised fund transfer Fraud and impersonation Not yet checked
2026-05-19Homebuyers lose $66,000 in suspected voice-cloning scam Fraud and impersonation Not yet checked
2026-05-13Yankton Boys & Girls Club deepfake bank fraud Fraud and impersonation Not yet checked
2026-05-06Deepfake video duped Belagavi man of Rs 7.9 lakh Fraud and impersonation Not yet checked
2026-03-11Researchers trick Perplexity Comet browser into phishing scam
Perplexity
Fraud and impersonation Not yet checked
2026-02-26Deepfake scam abuses Bitdefender name with Romanian officials' likenesses Fraud and impersonation Not yet checked
2026-02-10Police charge nine more in $34m deepfake crypto scam Fraud and impersonation Not yet checked
2026-02-02Family duped by AI video showing missing son kidnapped Fraud and impersonation Not yet checked
2025-12-15Deepfake Elon Musk car giveaway scam defrauds Florida man Fraud and impersonation Not yet checked
2025-12-09Grandparent scam with suspected AI voice cloning, Regina and Saskatoon (Saskatchewan, Canada, 2025) Fraud and impersonation Not yet checked
2024-08-01Deepfake liveness-check bypass attempts at Indonesian bank's digital KYC Fraud and impersonation Not yet checked

Also under this risk as a secondary risk: LR-2026-0750, LR-2026-0707, LR-2026-0672, LR-2026-0671, LR-2026-0673, LR-2026-0586, LR-2026-0094, LR-2026-0427, LR-2026-0352, LR-2026-0349, LR-2026-0318