Truepic brings two new spaces to Hugging Face, the first making C2PA Content Credentials available for developer use and the second an experimental space combining Content Credentials with Steg.AI’s digital watermarking algorithms.
San Diego, CA, Oct. 05, 2023 (GLOBE NEWSWIRE) — Truepic, a leading provider of authenticity infrastructure for the internet, announced today the launch of two innovative functions for generative AI users and developers. Through a partnership with Hugging Face, an open-access AI platform for hosting machine learning (ML) models, and Steg.AI, a provider of sophisticated forensic watermarking solutions, Hugging Face users can now automatically add responsible provenance metadata to AI-generated images.
The first Space, “Gen AI with Content Credentials,” demonstrates how generative AI platforms can empower creators to automatically and securely disclose their content as computer-generated at the point of creation. The demonstration utilizes several text-to-image models hosted on Hugging Face and combines them with Truepic’s technology to secure metadata into images in accordance with the Coalition for Content Provenance and Authenticity’s (C2PA) Content Credentials. That information is then displayed in an overlay on top of the image. The user can also transfer the media to compliant editing tools or a website that uses the C2PA open standard. Integrating the C2PA open standard into AI generation enables viewers to understand which platform content is created on while also ensuring it can travel throughout the internet while maintaining transparency as to its origin and history.
The second space, “Watermarked Content Credentials” represents the first public showcase of how emerging potential solutions to transparency — provenance and watermarking — can be combined. This approach adds the C2PA’s Content Credentials and Steg.AI’s watermarking technology to all image outputs simultaneously. The result is a significantly more robust layer of protection than would otherwise be possible. This combined method can help images better endure transforms, compression, and other real-world edits without losing their encoded provenance information. Both Truepic and Steg.AI are proud members of the C2PA.
“The emergence of high quality, AI-generated content has made it difficult to tell the difference between human and machine-created media,” said Nick Brown, EVP of Technology at Truepic. “It becomes particularly crucial when an image shows something that never actually happened or tricks you into thinking it’s real when it’s not. Responsible disclosure will create a more trustworthy information ecosystem where AI-generated content can thrive online. We are fortunate to partner with Hugging Face to showcase a higher level of transparency and visibility in marking and displaying AI-generated content with the C2PA’s Content Credentials. We are also excited to examine how Content Credentials and watermarking can be combined with Steg. AI on Hugging Face’s platform. These two emerging approaches have gained significant attention, and we are motivated to lead the way in examining the synergy between the two.”
“As generative AI is seeing growing adoption and helping create images that are becoming more difficult to distinguish from their training data, helping users better manage and annotate these new types of media will be essential to fostering more responsible and reliable development of the technology,” said Yacine Jernite, Machine Learning and Society Lead at Hugging Face. “We are excited to collaborate with Truepic to advance transparency and openness to better identify AI-generated content.”
“Seeing is no longer believing. In a world that’s become rife with image leaks, manipulation, and general misuse, the broad need for distinguishing trustworthy content has become clear. Attempting to recognize AI-generated imagery is a losing strategy. We’re seeing broad support for watermarking AI-generated content from US and EU government leaders and private industry,” said Dr. Eric Wengrowski, Co-Founder & CEO of Steg.AI. “We are excited to be partnered with Truepic and Hugging Face to deliver the first publicly usable forensic watermarking solution in a package that truly raises the bar for responsible AI. We acknowledge that this is uncharted waters, and so we’re using this demo to inform designs of future content attribution solutions.”
Hugging Face is a platform that prioritizes fostering a responsible environment, empowering diverse communities to assess the social implications of AI, and guiding the ethical development of ML models. This philosophy complements Truepic’s mission of creating positive change by establishing transparency as the foundation of all digital decision-making. These ethics support Steg.AI’s mission of building responsible AI technology for better security, privacy, and well-being.
Truepic provides authenticity infrastructure for the internet. The company’s transparency tools deliver verifiable information about the origin and editing history of digital content. Truepic is dedicated to advancing products and services that establish transparency in digital content, empowering viewers and users to identify authenticity and differentiate human from computer-generated. Recognized as one of TIME’s Best Inventions and Fast Company’s World Changing Ideas, consumers, businesses, creators, and publishers trust Truepic’s technologies to make informed decisions. For more information, visit truepic.com.
About Hugging Face
Hugging Face is the leading open-source and community-driven AI platform, providing tools that enable users to build, explore, deploy and train machine learning models and datasets. For more information, visit: huggingface.co.
Steg.AI builds state-of-the-art forensic watermarking software. Leading brands trust the company’s AI-powered watermarks to protect and authenticate their digital libraries. Steg.AI is venture-backed and supported by the United States National Science Foundation (NSF). For more information, visit steg.ai.