- Samsung has adopted the Content Credentials standard for AI-edited imagery
- The Samsung Galaxy S25 range will be the first phones in the world operating the standard
- The standard adds a tag and metadata to AI-edited images created on a Galaxy S25 device
Most of the attention at Samsung Galaxy Unpacked was obviously being devoted to the new phones – those in the Samsung Galaxy S25 range – but there was something quite important that slipped under the radar, and that’s the adoption of Content Credentials.
In 2024, the adoption of a standard for marking the creation of imagery and digital content was a hot topic, particularly due to the rise of generative AI and the plague of art theft that ensued to train large language models. Tech companies began adopting their own metadata markers and watermarks to signify AI altering, but a standard for identifying the legitimacy of an image has often been skipped.
One of the front runners for such a standard is Content Credentials, backed by the Content Authenticity Initiative (CAI). The tool is developed by Adobe and the Initiative counts Microsoft, Getty Images and Nvidia as members to name a few.
With this announcement, Samsung has joined the Coalition for Content Provenance and Authenticity (C2PA), which unifies the work of the CAI and its Content Credentials standard with Project Origin, another organization combatting misinformation but anchored in a news ecosystem that can verify the authenticity of content.
“We are excited to share that Samsung will implement #ContentCredentials for AI-generated images on the #GalaxyS25!” the C2PA wrote on LinkedIn. “Samsung has committed to a consequential step in bringing transparency to the digital ecosystem.”
If you suspect that an image has been altered with AI, then you can drop it into a tool built by Adobe to check its authenticity.
Think of Content Credentials as a ledger that contains content information; what device it has been captured on, what program (or AI tool) it has been altered with, even what settings were activated when the original image was created.
With this standard in tow, AI-generated and AI-altered images produced on Samsung Galaxy S handsets will receive a metadata-based label, basically noting that AI has tampered with what you’re seeing. The ‘CR’ watermark will also be added to the image. While the S25 family is the very first set of phones to carry the metadata marking on images, it follows camera companies Nikon and Leica who have also signed up to the standard.
The standard is, speaking broadly, a win for creatives looking to protect their work, but the obvious problem with any standard is a lack of enthusiasm. If not enough companies producing AI tools adopt standards that allow AI-altered content to be easily flagged, then such a system is worthless.
With more than 4,000 members under the wing of the Content Authenticity Initiative, here’s hoping tools to effectively flag the use of AI keep pace with the increasing capabilities of such tools.
Leave a comment