A New Approach to CSAM Detection
Method
Known CSAM
Unknown CSAM
Hash
10 - 30%
0%
L1ght AI
85%
85%
L1ght AI + Moderation
97%
97%
AVAILABLE NOW FOR: SOCIAL PLATFORMS, ADULT SITES, BROWSERS, SEARCH ENGINES
Contact us for integration details and to schedule a demo.
The spread of Child Sexual Abuse Material (CSAM) is a leading concen for Trust & Safety teams working to eliminate toxic content from their platforms and uphold company policies. The most common practice for identifying CSAM has been using hashes to compare against a central database of known CSAM.
L1ght brings a new and thorough approach that lets you act on CSAM immediately without waiting for someone else to report it.
L1ght’s CSAM detection technology is currently used by US law enforcement:
The Challenges
With Hashes
A hashing algorithm represents an image, video, or another form of data that can be used to determine if two files are identical or similar.
Similar to fingerprints, hashes are unique to the image they represent. This creates challenges for identifying CSAM as any slight change to the content will prevent the hash from matching, for example:
- Cropping or scaling the image
- Lightening or darkening the image
- Changing formats (jpeg, gif)
- Vertical or horizontal flip
L1ght’s Machine
Learning
L1ght’s anti-toxicity AI helps protect sites, apps and platforms from CSAM by continuously analyzing each image or video by running state of the art Machine Learning (ML) micro-classifiers.
- The actual content of the image or video is analyzed. This allows to detect CSAM content in both known (hashed) or unknown (new) content.
- Changes to the content do not affect the detection.
- 99+ sexual content classifiers are aggregated to provide detection of CSAM by age group.
Introducing L1ghtning:
The Trust & Safety API
Create AI Guardrails to Curb Online Toxicity on Your Platform.
Implement brand-protection, reduce user churn, and enhance trust & safety with 90+ models that analyze Text, Photos, Videos.
L1ghtning For Web Hosting Providers
Scan your hosted sites for problematic and illegal content, images and videos, including: Adult, Hate & CSAM.
Our Work Saves Children.
L1ght’s analysts, engineers and data scientists continuously work with technology companies and law enforcement to identify and bring child predators to justice.
L1ght Helps Facebook Remove 120,000 Child Predators from WhatsApp
Front-page story on the Financials Times
L1ght Helps Bing Remove CSAM from Search Results
It started with an anonymous tip to TechCrunch
Latest News
Doing Good Is No Longer Just For Nonprofits — It’s Also Tech Companies’ Responsibility
The L1ghtning API:
Taxonomic Classification
Web Services
Learn how we combine our product philosophy with our approach to data science as we create and expose an ever growing, breathing taxonomy of Online Toxicity.
Integrating anti-toxicity AI should be as easy as integrating a payments API.
Use our structured API or customize our 90+ models to identify, flag & respond to:
Bullying, Harassment, Predatory Behavior, Hate, Self-Harm, Grooming, CSAM.
NCMEC & Law Enforcement Reporting Included.
L1ghtning Strikes
Image & Video Analysis
L1ghtning is used to analyze millions of images per month for both known & un-hashed CSAM.
Website Analysis
Millions of websites are being scanned every month for CSAM, adult and illegal content using L1ghtning.
Messaging Analysis
L1ghtning integration scans in-app messaging for abusive language and behavior.
