Study: in May 2023, ~150K nonconsensual porn deepfakes, with 3.8B total views, appeared on 30 sites, up 9x since 2019, aided by services from big tech companies
To stay up and running, deepfake creators rely on products and services from Google, Apple, Amazon, CloudFlare and Microsoft X: @daveyalba , @cecianasta , and @ajmartinny X: Davey Alba / @daveyalba : NEW: Maybe you think of deepfake pornography as something that happens on the internet's edges. Not so. It's facilitated by Google, Microsoft, the app stores, hosting services and payment processors—all on the open web. @cecianasta and I document how: https://www.bloomberg.com/... Cecilia D'Anastasio / @cecianasta : NEW: Our months-long investigation reveals how big tech companies buoy and grow the deepfake AI porn industry that victimizes women. Google is the single largest traffic-driver to these sites. Amazon, Microsoft (GitHub) and Cloudflare provide crucial hosting services. Patreon,... [image] Andrew Martin / @ajmartinny : Excellent reporting here by @cecianasta and @daveyalba...... Google, Microsoft Tools Behind Surge in Deepfake AI Porn - https://www.bloomberg.com/...
Context & Ripple Effects
Earlier coverage described a nonconsensual-deepfake economy that was readily discoverable through search and, in some cases, supported by mainstream payment rails. This investigation extends that picture from individual sites to the infrastructure layers that keep them reachable and operating.
The issue had already moved beyond fringe porn-site distribution: deepfake pornography had become widespread on porn sites, while search visibility and card payments made the ecosystem easy to access. The study identifies the scale at which those dependencies matter.
First-order effects
- Google, Microsoft/GitHub, Amazon, Apple and Cloudflare become identifiable operational dependencies for an ecosystem whose creators and operators need discovery, hosting, app distribution and edge services.
- Google’s role as the largest traffic driver makes search distribution a particularly immediate point of exposure: removal or ranking changes could materially affect site acquisition.
Second-order effects
- The findings concentrate pressure on infrastructure providers to treat search, hosting, developer tools, app-store access and network services as connected enforcement surfaces rather than separate products.
- Operators facing restrictions at one layer may seek substitutes elsewhere, making consistent policies and cross-provider coordination more consequential than action by any single service.
Third-order effects
- If this pattern persists, responsibility for synthetic sexual-abuse content will increasingly be contested at the digital chokepoints that monetize, distribute and keep services online—not only at the sites publishing it.
- The broader challenge is an enforcement-surface problem: generative tools can expand supply quickly, while meaningful mitigation depends on whether intermediary controls can keep pace across a fragmented stack.
The trend: Nonconsensual deepfake pornography is becoming an infrastructure-governance issue as its distribution depends on mainstream discovery, hosting and platform services.