Microsoft Technology Center Against Child Sexual Abuse Material

Microsoft Technology Center, content filtering, image identification

Discover how Microsoft Technology Center is at the forefront of combatting the distribution of CSAM with PhotoDNA.

  • PhotoDNA calculates the unique characteristics of a digital image and creates a hash value to match even altered or edited images.
  • Microsoft Research and Dartmouth College provided it free to qualified organizations, including leading technology companies and law enforcement agencies.

In the first draft of this article, I beat around the bush. It was to make it more palatable for you, dear reader. But the subject isn’t, never will be, and should never be an easy one to write, read, or hear.  According to the Internet Watch Foundation, in 2022, they investigated a total of 375,230 reports possibly containing child sexual abuse material (CSAM), of which 255,580 reports did actually contain images or videos of children suffering sexual abuse. That’s about 68%. It’s horrifying. To combat this, in 2009, Microsoft Technology Center, in partnership with Dartmouth College, developed PhotoDNA, a proprietary image-identification and content-filtering technology.

What is It?        

PhotoDNA calculates the unique characteristics of a digital image and creates a hash value, or a digital fingerprint, for that image. Unlike other forms of hashing technology, PhotoDNA can still match images even if someone altered or edited them, making it groundbreaking in combating the distribution of CSAM.

Leading technology companies and law enforcement agencies have adopted tthis powerful technology.. Microsoft’s generosity extends to providing PhotoDNA to qualified organizations, including technology companies, developers, and non-profit organizations, free of charge. In 2015, Microsoft Technology Center made PhotoDNA available as a service on Azure, enabling smaller companies to harness its capabilities.

Okay… But How Does It Work?

PhotoDNA creates a unique hash value for each image. It then compares it to a database of known hashes to find identical or similar images. The hashing method used is resistant to alterations in the image, such as resizing or minor color changes. It does not rely on facial recognition and does not identify individuals or objects in an image. The technology ensures user privacy as it does not store any information about the photos themselves. The hash value is irreversible and Noone can create the original based on it.

Let’s say there’s a certain image that depicts sexual abuse toward a minor. PhotoDNA will then create hashes that specifically ID this one picture regardless of any edits made to the original image. The system will flag any image that has those hashes.

Who Uses It?

The National Center for Missing & Exploited Children (NCMEC) and major technology companies like Facebook, Twitter, Google, Adobe Systems, Reddit, and Discord utilize PhotoDNA to identify and remove CSAM from their platforms. NCMEC assigns PhotoDNA signatures to known abusive images, which are then shared with online service providers. These providers compare the signatures against the hashes of photos on their own services, enabling them to remove any matching images. Google alone has filed hundreds of thousands of reports of CSAM and disabled thousands of user accounts using this technology.

But…

While PhotoDNA is an effective tool in combating the distribution of known abusive images, it relies on pre-identified images with known hashes. This means that it cannot detect new or previously unseen abusive images.

Final Thoughts

PhotoDNA, with the support and contribution from the Microsoft Technology Center, is an invaluable technology in the fight against child exploitation and the distribution of abusive images online. When I first looked into it, I was surprised that not only is a Big Tech company making something so noble but also it’s not monetizing it. Do with that as you will.


Inside Telecom provides you with an extensive list of content covering all aspects of the tech industry. Keep an eye on our Intelligent Tech sections to stay informed and up-to-date with our daily articles.