Tech News

What are content credentials? Here’s why Adobe’s new AI keeps this metadata front and center



Content Credentials Website

Adobe has been at the forefront of generative AI, unveiling numerous AI solutions, from your classic text-to-image generator to tools specifically designed for businesses. Regardless of the application, one thing remains the same: the presence of Content Credentials.

At the Adobe Summit this week, the company extended Content Credentials to its Creative and Enterprise apps, including its new framework credential in Firefly and Adobe Experience Manager.

Also: This new AI tool from Adobe makes it even easier to generate the images you need.

This addition highlights Adobe’s continued commitment to helping audiences differentiate between authentic and fake content – a problem exacerbated by the growing presence of generative AI tools – by including tamper-proof metadata that includes information about how the content was created.

“The technology problem here is the easiest to solve – we’ve solved it; adoption is the second hardest thing to solve, but we’re showing some traction,” said Andy Parsons, senior director of Content Authenticity Initiative at Adobe, at ZDNET. .

At the summit, Adobe announced that PricewaterhouseCoopers (PwC), NBCUniversal News Group, WPP and Autodesk have joined the Content Authenticity Initiative, an organization that promotes “the adoption of an open industry standard for content authenticity and provenance “. Over the past year, tech giants such as OpenAI, Meta, and Google have supported the use of content credentials.

Also: Do ​​you want to work in AI? How to pivot your career in 5 steps

“You see these big brands coming on board to not only give us their logo and say this is a good idea, but also to implement content credentials,” Parsons added.

What are content credentials?

So why are these companies all about content credentials? Let’s look at content credentials and how you can use them.

Content IDs serve as a “nutrition label” for digital content, allowing users to see the essential ingredients of the content, including creator name, time and date of creation, tools used to create the content (such as a camera or software) and the changes made. .

The information is permanently added to the content via cryptographic metadata and a watermark, ensuring that the identifying information remains attached, even if someone modifies the data or content.

Also: The Best AI Image Generators of 2024: Tested and Reviewed

Since Content Credentials are based on free and open source technology, creators can choose to attach Content Credentials to any content to ensure proper credit is given. Additionally, many AI applications, such as Adobe Firefly, automatically implement content credentials to generate content.

Therefore, it is significant that major players in the generative AI space, such as OpenAI and Meta, have shown their commitment to using content credentials to label images generated in their content models. AI, thus working towards a world in which all AI-generated images are labeled. correctly.

Why labeling AI-generated images is so important

As generative AI tools become more popular and creating realistic fake images becomes easier than ever, the need to correctly identify AI-generated content becomes more critical.

“As (generative AI) tools become more democratized, become low-cost or free, and the quality of these low-cost or free tools produced becomes indistinguishable from the real thing, of course, (credentials content are) increasingly important,” Parsons said.

Also: We are not ready to deal with the impact of generative AI on elections

Malicious actors can use generative AI to create false content to spread misinformation and mislead the public. This trend poses a particularly imminent threat in an election year.

“A real concern (is) that these elections this year in 2024 may be election season for AI, where people’s understanding of what is real, what the candidates are doing, what “They don’t… (there) will be enough misleading content (which will add) a lot of noise to the system,” Parsons added.

To combat this threat, and in addition to content credentials, Adobe Firefly is taking other steps to ensure that AI generations of politicians or celebrities are not possible.

Also: 6 AI tools to boost your work and daily life

Adobe AI was trained only on Adobe Stock images and public domain content with expired copyrights. The technology cannot generate an image of a celebrity or political candidate because it was not trained on that data.

“So between these checks, the ability to add this metadata and the way the image model is trained on openly licensed content – we’re not scraping the open internet – the Firefly image model has no “I couldn’t create a Mickey Mouse or a Donald Trump because it’s never essentially about seeing a Mickey Mouse or a Donald Trump,” Deepa Subramaniam, Adobe’s vice president of product marketing, told ZDNET.

Even for applications that do not prohibit public persona generation and are trained by Internet scraping, the presence of content credentials can help solve the problem of content provenance and authenticity by allowing users to easily check the source of an image. , without knowing anything about metadata.

How to use content credentials

For the consumer, it’s easy to get started with Content Credentials. If you see the Content Credentials pin next to a piece of content, you can click it to view all of the image’s most vital information, as shown in the image at the top of the article . When a Pin is not available, you can upload the image to the Content Credentials website.

Once the photo is uploaded, if there is any content identifying information in the metadata, it will be presented to you. Otherwise, it will match your image to similar images on the internet and tell you whether those images were AI-generated or not.

Also read: Adobe announces generative AI tools to reinvent advertising campaigns

I tried the technology by capturing the viral image of Pope Francis wearing a down jacket and uploading it to the Content Credentials website. Since it was a screenshot, the image did not contain metadata. However, the site compared the photo to other images found on the internet and identified that it was generated by AI, as shown in the image below:

Content identification information

Screenshot by Sabrina Ortiz/ZDNET

Using content credentials was intuitive, simple, and helpful in finding out what I wanted to know, and it will become easier over time as the content credentials pin becomes more prevalent on the internet.

“I think the best way as a consumer to learn about things is to see them in action: touch them, click them, see how they work,” Parsons said. “So all of these resources are available at”

Disclosure: The cost of Sabrina Ortiz’s trip to Las Vegas for the Adobe Summit was covered by Adobe, a standard industry practice for long-distance travel. The judgments and opinions of ZDNET writers and editors are always independent of the companies we cover.


Related Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button