Demo used in the Google blog post showing an example of how a Midjourney-generated image might look in a Google search results panel.
Mockup shown in the Google blog post depicting an example of how a Midjourney-generated image might look in a Google search results panel.

As a follow-up to yesterday’s news on Google using IPTC metadata to mark AI-generated content we are happy to announce that generative AI tools from Midjourney and Shutterstock will both be adopting the same guidelines.

According to a post on Google’s blog, Midjourney and Shutterstock will be using the same mechanism as Google – that is, using the IPTC “Digital Source Type” property to embed a marker that the content was created by a generative AI tool. Google will be detecting this metadata and using it to show a signal in search results that the content has been AI-generated.

A step towards implementing responsible practices for AI

We at IPTC are very excited to see this concrete implementation of our guidance on metadata for synthetic media.

We also see it as a real-world implementation of the guidelines on Responsible Practices for Synthetic Media from the Partnership on AI, and of the AI Ethical Guidelines for the Re-Use and Production of Visual Content from CEPIC, the alliance of European picture agencies. Both of these best practice guidelines emphasise the need for transparency in declaring content that was created using AI tools.

The phrase from the CEPIC transparency guidelines is “Inform users that the media or content is synthetic, through
labelling or cryptographic means, when the media created includes synthetic elements.”

The equivalent recommendation from the Partnership on AI guidelines is called indirect disclosure:

“Indirect disclosure is embedded and includes, but is not limited to, applying cryptographic provenance to synthetic outputs (such as the C2PA standard), applying traceable elements to training data and outputs, synthetic media file metadata, synthetic media pixel composition, and single-frame disclosure statements in videos”

Here is a simple, concrete way of implementing these disclosure / transparency guidelines using existing metadata standards.

Moving towards a provenance ecosystem

IPTC is also involved in efforts to embed transparency and provenance metadata in a way that can be protected using cryptography: C2PA, the Content Authenticity Initiative, and Project Origin.

C2PA provides a way of declaring the same “Digital Source Type” information in a more robust way, that can provide mechanisms to retrieve metadata even after the image was manipulated or after the metadata was stripped from the file.

However implementing C2PA technology is more complicated, and involves obtaining and managing digital certificates, among other things. Also C2PA technology has not been implemented by platforms or search engines on the display side.

In the short term, AI content creation systems can use this simple mechanism to add disclosure information to their content.

The IPTC is happy to help any other parties to implement these metadata signals: please contact IPTC via the Contact Us form.

Sundar Pichai, CEO of Google, extolling the benefits of image metadata at Google IO 2023.

At today’s Google I/O event keynote, Sundar Pichai, CEO of Google, explained how Google will be using embedded IPTC image metadata to signal visual media created by generative AI models.

“Moving forward, we are building our models to include watermarking and other techniques from the start,” Pichai said. “If you look at a synthetic image, it’s impressive how real it looks, so you can imagine how important this is going to be in the future.

“Metadata allows content creators to associate additional context with original files, giving you more information whenever you encounter an image. We’ll ensure every one of our AI-generated images has that metadata.”

The IPTC Photo Metadata section of Google Images’ guidance on metadata has been updated with new guidance on the DigitalSourceType field:

This follows the guidance on IPTC Photo Metadata for Generative AI that was recently published by IPTC.

“AI-Generated” label on Google Images

The above guidance hints at an “AI-generated label” to be used on Google Images in the future. Google recommends that all creators of AI-generated images use the IPTC Digital Source Type property to signal AI-generated content. While Google says that “you may not see the label in Google Images right away”, it appears that it will soon be available in Google Images search results.

AI-generated image of a cute robot sitting at a garden table sketching on a notepad.
Image created by Brendan Quinn using Bing Image Creator. This image file contains digitalsourcetype metadata which was added manually using exiftool.

The IPTC has updated its Photo Metadata User Guide to include some best practice guidelines for how to use embedded metadata to signal “synthetic media” content that was created by generative AI systems.

After our work in 2022 and the draft vocabulary to support synthetic media, the IPTC NewsCodes Working Group, Video Metadata Working Group and Photo Metadata Working Group worked together with several experts and organisations to come up with a definitive list of “digital source types” that includes various types of machine-generated content, or hybrid human and machine-generated media.

Since publishing the vocabulary, the work has been picked up by the Coalition for Content Provenance and Authenticity (C2PA) via the use of digitalSourceType in Actions and in the IPTC Photo and Video Metadata assertion. But the primary use case is for adding metadata to image and video files

Here is a direct link to the new section on Guidance for using Digital Source Type, including examples for how the various terms can be used to describe media created in different formats – audio, video, images and even text.

IPTC recommends that software creating images using trained AI algorithms uses the “Digital Source Type” value of “trainedAlgorithmicMedia” is added to the XMP data packet in generated image and video files. Alternatively, it may be included in a C2PA manifest as described in the IPTC assertion documentation in the C2PA specification.

The official URL for the full vocabulary is http://cv.iptc.org/newscodes/digitalsourcetype, so the complete URI for the recommended Trained Algorithmic Media term is http://cv.iptc.org/newscodes/digitalsourcetype/trainedAlgorithmicMedia.

Other terms in the vocabulary include:

Of course, the original digital source type values covering photographs taken on a digital camera or phone (digitalCapture), scan from negative (negativeFilm),  and images digitised from print (print) are also valid and may continue to be used. We have, however, retired the generic term “softwareImage” which is now deemed to be too generic. We recommend using one of the newer terms in its place.

If you are considering implementing this guidance in AI image generation software, we would love to hear about it so we can offer advice and tell others. Please contact us using the IPTC contact form.

DALL-E image: "An abstract painting of new year's fireworks in the sky, over an sea made of electronic circuit boards"
Image generated by DALL-E, based on the prompt: “An abstract painting of new year’s fireworks in the sky, over an sea made of electronic circuit boards”

Here is a wrap-up of IPTC has been up to in 2022, covering our latest work, including updates to most of our key standards.

Two successful member meetings and five member webinars

This year we again held our member meetings online, in May and October. We had over 70 registered attendees each time, from over 40 organisations, which is well over half of our member organisations so it shows that the virtual format works well.

This year we had guests from United Robots, Kairntech, EDRLab, AxateHAND Identity, RealityDefender.ai, synthetic media consultant Henrik de Gyor and metaverse expert Toby Allen, as well as member presentations from The New York Times, Agence France-Presse, Refinitiv (an LSE Group company), DATAGROUP ConsultingTT Sweden, iMatrics and more. And that’s not even counting our regular Working Group presentations! So we had a very busy three days in May and October.

We also had some very interesting members-only webinars including a deep dive into ninjs 2.0, JournalList and the trust.txt protocol, a joint webinar with the EBU on how Wikidata and IPTC Media Topics can be used together, and a great behind the scenes question-and-answer session with a product manager from Wikidata itself.

Recordings of all presentations and webinars are available to IPTC members in the Members-Only Zone.

A fascinating Photo Metadata Conference

This year’s IPTC Photo Metadata Conference was held online in November and we had over 150 registrants and 19 speakers from Microsoft, CBC Radio Canada, BBC, Adobe, Content Authenticity Initiative, the Smithsonian and more. The general theme was bringing the IPTC Photo Metadata Standard to the real world, focussing on adoption of the recently-introduced accessibility properties, looking at adoption and interoperability between different software tools, including a new comparison tool that we have introduced; use of C2PA and Content Authenticity in newsroom workflows, with demos from the BBC and CBC (with Microsoft Azure).

We also had an interesting session discussing the future of AI-generated images and how metadata could help to identify which images are synthetic, the directions and algorithms used to create them, and whether or not the models were trained on copyrighted images.

Recordings of all sessions are available online.

Presentations at other conferences, work with other organisations

IPTC was represented at the CEPIC Congress in Spain, the DigiTIPS conference run by imaging.org, the Sports Video Group’s content management group, and several Project Origin events.

Our work with C2PA is progressing well. As of version 1.2 of the C2PA Specification, assertions can now include any property from IPTC Photo Metadata Standard and/or IPTC Video Metadata Hub. C2PA support is growing in tools and is now available in Adobe Photoshop.

IPTC is also working with Project Origin on enabling C2PA in the news industry.

We had an IPTC member meet-up at the NAB Show in Las Vegas in May.

We also meet regularly with Google, schema.org, CIPA (the camera-makers behind the Exif standard), ISO, CEPIC and more.

Standard and Working Group updates

  • Our IPTC NewsCodes vocabularies had regular updates each quarter, including 12 new terms at least 20 retired terms. See the details in our news posts about the September Update, July Update, May Update, and the February Update (in time for the Winter Olympics). We also extended the Digital Source Type vocabulary specifically to address “synthetic media” or AI-generated content.
  • The News in JSON Working Group released ninjs 1.4, a parallel release for those who can’t upgrade to ninjs 2.0 which was released in 2021. We published a case study showing how Alamy uses ninjs 2.0 for its content API.
  • NewsML-G2 v2.31 includes support for financial instruments without the need to attach them to organisations.
  • Photo Metadata Standard 2022.1 includes a Contributor structure aligned with Video Metadata Hub which can handle people who worked on a photograph but did not press the shutter, such as make-up artists, stylists or set designers;
  • The Sports Content Working Group is working on the IPTC Sport Schema, which is pre-release but we are showing it to various stakeholders before a wider release for feedback. If you are interested, please let me know!
  • Video Metadata Hub 1.4 includes new properties for accessibility, content warnings, AI-generated content, and clarifies the meanings of many other properties.

New faces at IPTC

We waved farewell to Johan Lindgren of TT as a Board Member, after five years of service. Thankfully Johan is staying on as Lead of the News in JSON Working Group.

We welcomed long-time member Heather Edwards of The Associated Press as our newest board member.

We welcomed Activo, Data Language, Denise Kremer, MarkLogic, Truefy, Broadcast Solutions and Access Intelligence as new IPTC members, plus Swedish publisher Bonnier News who are joining at the start of 2023. We’re very happy to have you all as members!

If you are interested in joining, please fill out our membership enquiry form.

Web site updates

We launched a new, comprehensive navigation bar on this website, making it easier to find our most important content.

We have also just launched a new section highlighting the “themes” that IPTC is watching across all of our Working Groups:

We would love to hear what you think about the new sections, which hopefully bring the site to life.

Best wishes to all for a successful 2023!

Thanks to everyone who has supported IPTC this year, whether as members, speakers at our events, contributors to our standards development or software vendors implementing our standards. Thanks for all your support, and we look forward to working with you more in the coming year.

If you have any questions or comments, you can contact me directly at mdirector@iptc.org.

Best wishes,

Brendan Quinn
Managing Director, IPTC

We had a great Photo Metadata Conference last Thursday. Thanks to those who attended. For those who didn’t, or those who would like to go over some detail again, here we publish full recordings of all sessions.

First up, Brendan Quinn, IPTC Managing Director introduced the day and gave an overview of what was to come:

Next was a great panel on adoption of the accessibility properties added in the 2021.1 update to the IPTC Photo Metadata Standard. We are very happy to share that the fields are now supported in many popular photo creating and editing tools, with more to come:

Next was David Riecks and Michael Steidl, co-leads of the IPTC Photo Metadata Working Group, presenting the work done by the Working Group since the last Photo Metadata Conference:

Then came a session on real-world implementations of the C2PA specification for content authenticity, including presentations from Microsoft, CBC / Radio Canada, the BBC and Adobe / Content Authenticity Initiative:

 

The last session was a panel discussion on Metadata for AI Images, looking at questions around the ethics of using copyrighted content to train a machine learning engine to generate AI images, and how the IPTC Photo Metadata Standard could be extended to support metadata appropriate for AI-generated images:

We had a great session and a packed conference! We look forward to seeing everyone again at next year’s event.

Screenshot of the standard specification for the IPTC Photo Metadata Standard 2022.1 version.Today, the IPTC announces the release of the IPTC Photo Metadata Standard, version 2022.1.

The update has some changes to align the IPTC Photo Metadata Standard more closely which IPTC Video Metadata Hub, which helps with the integration into the C2PA specification as announced earlier this week.

The changes will be presented today (Thursday 10 November) at the IPTC Photo Metadata Conference, alongside sessions on AI and images, C2PA, and accessibility. You can still register for today’s Photo Metadata Conference, for free, at the IPTC Photo Metadata Conference 2022 event page. Video recordings from the event will be posted in the coming weeks.

The changes in detail

The IPTC Core schema has been updated to version 1.4, including the following changes:

  • Name of property Source changed to Source (Supply Chain)

  • Property Subject Code was set to legacy state

The IPTC Extension schema has been updated to version 

  • New property Contributor (matching the equivalent property in IPTC Video Metadata Hub)

  • The property structure for Product has been extended with a new property Identifier

The specification document has also been updated with some edits and additions to Help Texts and User Notes.

Technical Reference update

The IPTC Photo Metadata TechReference has also been updated to include the changes in version 2022.1. This can be used by software makers to easily include the changes in the new version.

For more information on how to use the Technical Reference, please consult the IPTC Photo Metadata Technical Reference documentation.

Questions? Comments?

For any questions or comments on this update or on the IPTC Photo Metadata Standard in general, please post to the public IPTC Photo Metadata forum at iptc-photometadata@groups.io.

Screenshot of the section of the C2PA 1.2 specification showing the new IPTC assertion definition.
Screenshot of the section of the C2PA 1.2 specification showing the new IPTC assertion definition.

We are happy to announce that IPTC’s work with C2PA, the Coalition for Content Provenance and Authority, continues to bear fruit. The latest development is that C2PA assertions can now include properties from both the IPTC Photo Metadata Standard and our video metadata standard, IPTC Video Metadata Hub.

Version 1.2 of the C2PA Specification describes how metadata from either the photo or video standard can be added, using the XMP tag for each field in the JSON-LD markup for the assertion.

For IPTC Photo Metadata properties, the XMP tag name to be used is shown in the “XMP specs” row in the table describing each property in the Photo Metadata Standard specification. For Video Metadata Hub, the XMP tag can be found in the Video Metadata Hub properties table under the “XMP property” column.

We also show in the example assertion how the new accessibility properties can be added using the Alt Text (Accessibility) field which is available in Photo Metadata Standard and will soon be available in a new version of Video Metadata Hub.

Title slide of Sam Joehl's presentation "What does an Image Sounds LIke?" from the 2021 IPTC Photo Metadata Conference
Title slide of Sam Joehl’s presentation “What does an Image Sounds Like?” from the 2021 IPTC Photo Metadata Conference

The IPTC Photo Metadata Working Group is proud to announce the IPTC Photo Metadata Conference 2022. The event will be held online on Thursday November 10th from 15.00 – 18.00 UTC.

This year the theme is Photo Metadata in the Real World. After introducing two new developments last year: the IPTC Accessibility properties and the C2PA specification for embedding provenance data in photo and video content – we re-visit both technologies to see how they are being adopted by software systems, publishers and broadcasters around the world.

The 3-hour meeting will host four sessions:

  • Adoption of the IPTC Accessibility Properties – we hear from vendors and content creators on how they are progressing in implementing the new properties to support accessibility
  • Software Supporting the IPTC Photo Metadata Standard – showcasing an update to IPTC’s directory of software supporting the IPTC Photo Metadata Standard, including field-by-field reference tables letting users compare software implementations
  • Use of C2PA in real workflows – showcasing early work on implementing the C2PA specification in media organisations
  • Artificial Intelligence and metadata – looking at the questions around copyright and synthetic media: for example,  when generative AI uses thousands of potentially copyrighted images to train machine learning models, who owns the resulting images?

We look forward to welcoming all interested parties to the conference – no IPTC membership is needed to attend. The event will be held as a Zoom webinar.

Please see more information and the Zoom registration link on the event page.

See you there on the November 10th!

Chart showing the main players in the metaverse authoring and publishing value chain
A screenshot from Toby Allen’s talk on how media organisations can prepare for the Metaverse.

Last week the IPTC held its Autumn Meeting 2022, with over 70 attendees from over 20 countries attending the three-day online event.

Discussions were as wide-ranging as ever. Highlights were a guest presentation on how media organisations can prepare for the Metaverse from startup advisor and previous member of the Microsoft HoloLens team, Toby Allen; intense discussions from members about making our work in machine-readable rights and RightsML simpler and more accessible and to bridge the gap between our simple, lightweight JSON news standard ninjs and our richly structured full-featured XML-based standard NewsML-G2.

We also heard about many other topics:

  • Meinolf Ellers at IPTC member dpa spoke about the DRIVE initiative, which follows on from the C-POP project that IPTC advised on in 2019 and 2020. DRIVE allows consortium members to share data about content usage to drive subscriptions and engagement, and to find under-represented areas in their news output to meet audience needs.
  • We heard about representing social media content in NewsML-G2: Dave Compton of Refinitiv spoke about their work encoding content from Twitter and other social networks in NewsML-G2 format for re-use, enhancement and syndication.
  • Will Kreth, previously CEO of EIDR, spoke about the HAND project which aims to create a unique identifier for media and sports talent
  • Fredrik Lundberg from IPTC member iMatrics and guest presenter Jens Pehrson from GOTA Media spoke about a new tool they have developed that allows publishers to track the gender balance in their news content
  • Johan Lindgren from IPTC member TT (the Swedish national news agency) spoke about their recent project to develop a classification and entity extraction engine for their news content, based on IPTC Media Topics taxonomy
  • We heard from Audren Layeux of CARSA who spoke about the European Media Data Space project, an EU initiative
  • Ben Colman, CEO of RealityDefender spoke (direct from TechCrunch Disrupt in San Francisco!) about their deepfake detection technology, used by social media networks, financial institutions and media organisations to detect manipulated images and videos.
  • IPTC MD Brendan Quinn spoke about IPTC’s ongoing work with C2PA and Project Origin, including forthcoming additions to C2PA to include video metadata.

In addition, we heard updates from all IPTC Working Groups: Dave Compton introduced NewsML-G2 2.31; Paul Kelly spoke about some new developments in the RDF-based sports data model which will be announced soon; Pam Fisher described the work of the Video Metadata Working Group and the changes coming in Video Metadata Hub v1.4; David Riecks and Michael Steidl spoke about Photo Metadata Standard 2022.1 and the ongoing work of the Photo Metadata Working Group; 

The Standards Committee voted in new standard versions: NewsML-G2 v2.31, Video Metadata Hub v1.4, and Photo Metadata Standard 2022.1. These will be released and publicised over the coming weeks.

The IPTC Annual General Meeting 2022 saw Johan Lindgren step down from the Board of Directors after 6 years of service. Thanks very much for all your help, Johan!

We are very happy to welcome a new Board member: Heather Edwards of Associated Press.

Thanks very much to everyone who attended and spoke. You contributed to making it a great event for all!

As usual, full recordings of all sessions are available to IPTC members on the members-only event page.

Family Tree magazine has published a guide on using embedded metadata for photographs in genealogy – the study of family history.Screenshot of the beginning of the article on FamilyTree.com describing how to use IPTC photo metadata for genealogy

Rick Crume, a genealogy consultant and the article’s author, says IPTC metadata “can be extremely useful for savvy archivists […] IPTC standards can help future-proof your metadata. That data becomes part of the digital photo, contained inside the file and preserved for future software programs.”

Crume quotes Ken Watson from All About Digital Photos saying “[IPTC] is an internationally recognized standard, so your IPTC/XMP data will be viewable by someone 50 or 100 years from now. The same cannot be said for programs that use some proprietary labelling schemes.”

Crume then adds: “To put it another way: If you use photo software that abides by the IPTC/XMP standard, your labels and descriptive tags (keywords) should be readable by other programs that also follow the standard. For a list of photo software that supports IPTC Photo Metadata, visit the IPTC’s website.

“[IPTC] is an internationally recognized standard, so your IPTC/XMP data will be viewable by someone 50 or 100 years from now”

The article goes on to recommend particular software choices based on IPTC’s list of photo software that supports IPTC Photo Metadata. In particular, Crume recommends that users don’t switch from Picasa to Google Photos, because Google Photos does not support IPTC Photo Metadata in the same way. Instead, he recommends that users stick with Picasa for as long as possible, and then choose another photo management tool from the supported software list.

Similarly, Crume recommends that users should not move from Windows Photo Gallery to the Windows 10 Photos app, because the Photos app does not support IPTC embedded metadata.

Crume then goes on to investigate popular genealogy sites to examine their support for embedded metadata, something that we do not cover in our photo metadata support surveys.

The full article can be found on FamilyTree.com.