Image scraping bots systematically harvest photographs from public and semi-public sources — social media profiles, news articles, professional websites, and dating apps. The collected images are then used to train AI models for deepfake generation, face-swap tools, and nudify applications. A person's entire public photo history can be scraped in seconds, providing more than enough reference material for AI model fine-tuning. Most platforms prohibit scraping in their terms of service, and the Computer Fraud and Abuse Act may apply to unauthorized scraping. However, scraping of publicly visible content is difficult to prevent technically.

Key facts about this term

  1. Limiting public photos reduces scraping risk Setting social media profiles to private and minimizing public photo exposure reduces the amount of reference material available for AI model training.
  2. Platform terms of service prohibit scraping Major platforms prohibit automated data collection. Victims of scraping-based NCII may have additional civil claims against scrapers who violated terms of service.
  3. Scraping enables AI NCII without any prior relationship Unlike traditional NCII (which requires accessing images from a relationship), scraping allows strangers to target anyone with a public social media presence.

Frequently asked questions

Can I find out if my photos were scraped?

Direct detection of scraping is difficult. ScanErase's biometric scan identifies the output — AI-generated deepfake content featuring your face — even when the scraping itself cannot be detected.

Does making my profile private prevent scraping?

Making profiles private prevents future scraping. However, photos that were publicly accessible before privatization may already have been scraped. ScanErase can identify if deepfake content exists from those prior photos.

Related glossary terms