TAKE IT DOWN Act Nets First AI Deepfake Porn Arrests
Key insights
- Shannon and Hernandez collectively targeted 140+ victims across 473 albums that drew millions of combined views.
- The TAKE IT DOWN Act carries a maximum two-year federal prison sentence and passed with bipartisan congressional support.
- An Ohio man's guilty plea last month for AI-generated CSAM marked the act's first conviction, preceding these arrests by weeks.
Why this matters
Federal enforcement of AI-generated deepfakes has crossed from symbolic prosecution into active multi-defendant investigation, with two cases filed within months of the law's first guilty plea. The TAKE IT DOWN Act now gives prosecutors a federal framework deployable against high-volume image abuse operations, not just edge-case offenders, which changes the risk calculus for platforms hosting user-generated AI content. AI practitioners building or distributing image generation tools face regulatory and liability scrutiny that extends beyond their own platforms to downstream applications generating non-consensual intimate imagery at scale.
Summary
The DOJ and FBI made their first arrests under the TAKE IT DOWN Act, charging Cornelius Shannon (51) and Arturo Hernandez (20) for publishing AI-generated deepfake pornography targeting 140+ female victims.
Shannon's 360 albums drew millions of views; Hernandez's 113 albums nearly a million. Both face up to two years in federal prison under legislation signed with bipartisan support last year.
Essentially: (DOJ, FBI) have moved from symbolic first conviction to active, multi-defendant enforcement of AI-generated image abuse.
- Shannon's 360-album volume suggests organized, systematic abuse rather than isolated incidents.
- Charges follow an Ohio man's TAKE IT DOWN Act guilty plea last month, covering AI-generated CSAM.
- The law covers both adult non-consensual intimate imagery and AI-generated CSAM under one federal statute.
Platforms that hosted millions of Shannon's views now face real enforcement pressure, not just policy obligations.
Potential risks and opportunities
Risks
- Image generation platform providers (Stability AI, Midjourney, RunwayML) face subpoenas or platform audits if their tools are identified in discovery as the source of Shannon's or Hernandez's deepfakes
- Social and content platforms that hosted millions of Shannon's views could face parallel civil suits from victims under state NCII statutes, which carry damages beyond the federal two-year cap
- The two-year sentencing ceiling may draw public criticism from victim advocacy groups as inadequate relative to abuse scale, creating legislative pressure to amend TAKE IT DOWN Act penalties in the current congressional session
Opportunities
- Content moderation and CSAM detection vendors (Thorn, ActiveFence, Hive Moderation) can now pitch proactive deepfake screening tools directly to platforms implicated by the scale of Shannon's hosting reach
- Digital forensics firms with AI-attribution and provenance-tracing capabilities gain direct leverage as prosecutors will require expert witnesses to link generated imagery back to specific models or tools in future TAKE IT DOWN Act cases
- Victim advocacy organizations and legal tech platforms offering non-consensual intimate imagery reporting tools see renewed federal and foundation funding momentum following the first high-profile multi-defendant enforcement action
What we don't know yet
- Whether the platforms hosting Shannon's millions-of-views albums face any takedown obligation or liability exposure under TAKE IT DOWN Act provisions
- Which specific AI generation tools Shannon and Hernandez used to produce the deepfakes, undisclosed in public reporting as of May 2026
- How investigators verified each of the 140+ victims' identities and whether all victims had been notified before charges were filed
Originally reported by fortune.com
Read the original article →Original headline: First Criminal Charges Filed Under TAKE IT DOWN Act — Two Men Arrested for Publishing AI Deepfake Pornography Depicting 140+ Victims