Home › Legal Risk
Enforcement briefing · November 2026
Statute counts, penalty ceilings and removal speeds — the legal environment for undress ai rendered as a dataset.
Between 2019 and 2026 the legal status of undress ai output moved from a patchwork of leaked-photo statutes to a purpose-built regime spanning every major jurisdiction. Counting operative instruments rather than headlines gives the clearest picture of that shift.
Counting jurisdictions with operative intimate-deepfake provisions by year shows the enforcement wave arriving later than the technology — then compounding:
| Jurisdiction | Offense | Ceiling | Notes |
|---|---|---|---|
| US — federal | Distribution (TAKE IT DOWN) | 2 yrs; 3 yrs if minor depicted | Plus platform 48-h removal duty |
| US — states | Distribution; creation (minority of states) | Misdemeanor → felony | ~48 + DC criminalize sharing; creation offenses expanding |
| United Kingdom | Sharing (OSA) · Creating (DUAA) | 2 yrs each | Only full lifecycle coverage in dataset |
| Australia | Sharing (Criminal Code 2024) | 6 yrs (7 aggravated) | Highest ceiling measured |
| EU member states | Sharing under national IBSA laws | ~1–3 yrs typical | DSA/GDPR duties stack on top |
| All jurisdictions | Depicting a minor | CSAM statutes — decades | Separate universe; no tolerance threshold |
| Rank | Route | Speed | Coverage | Cost |
|---|---|---|---|---|
| 1 | TAKE IT DOWN request (covered platforms) | ≤48 h — statutory | US-covered platforms | Free |
| 2 | StopNCII hash-match (adults) | ~24–72 h propagation | Partner network; blocks re-uploads | Free |
| 3 | NCMEC Take It Down (depictions from minority) | Days | Partner network | Free |
| 4 | Platform report / DMCA on source photo | Days–weeks | Broad but inconsistent | Free/minimal |
| 5 | Criminal complaint | Months; reaches the creator | Jurisdiction-dependent | Free |
| 6 | Civil action (§ 6851, state torts) | Months–years; damages | Identifiable defendants | Fees |
The operative pattern in the data: victims who act in the first 48–72 hours — hash plus statutory demand, in parallel — achieve materially better suppression outcomes than sequential filers. Every subsequent day the content spends unhashed measurably increases re-upload count. Scale is on the victim's side here: StopNCII's hash registry has accumulated hundreds of thousands of fingerprints since launch, and its partner network covers the platforms where the overwhelming majority of re-uploads occur — which is what puts hash-matching at rank 2 despite being free and device-side.
¹ State counts include statutes written for intimate imagery generally and applied to synthetics, plus newer deepfake-specific provisions.
² Australia's aggravated ceiling applies where conduct is connected to other offenses (e.g., harassment, minors-adjacent circumstances).
³ Cumulative count treats US states individually because exposure varies state by state; EU member states are aggregated under the framework count to avoid double-counting.