On Thursday, Google announced that it will be extending its usage of AI content labelled using the IPTC Digital Source Type vocabulary.

We have previously shared that Google uses IPTC Photo Metadata to signal AI-generated and AI-edited media, for example labelling images edited with the Magic Eraser tool on Pixel phones.

In a blog post published on Friday, John Fisher, Engineering Director for Google Photos and Google One posted that “[n]ow we’re taking it a step further, making this information visible alongside information like the file name, location and backup status in the Photos app.”

This is based on IPTC’s Digital Source Type vocabulary, which was updated a few weeks ago to include new terms such as “Multi-frame computational capture sampled from real life” and “Screen capture“.

Google already surfaces Digital Source Type information in search results via the “About this image” feature.

Also, the human-readable label for the term http://cv.iptc.org/newscodes/digitalsourcetype/trainedAlgorithmicMedia was clarified to be “Created using Generative AI” and similarly the label for the term  http://cv.iptc.org/newscodes/digitalsourcetype/compositeWithTrainedAlgorithmicMedia was clarified to be “Edited with Generative AI.” These terms are both used by Google.

The IPTC NewsCodes Working Group is pleased to announce the latest release of the IPTC NewsCodes, our set of controlled vocabularies for the news industry.

Updates this time span many vocabularies, with the biggest updates to Media Topic and Digital Source Type.

Media Topic updates

Most of the recent work has been in the politics branch.

3 new concepts: by-election, recall election, coalition building

2 retired concepts: political campaigns, church elections

4 modified concept names (in English): voting system, referendum, fundamental rights, football (yes we finally refer to the sport as “football” in en-GB and “soccer” in en-US!)

Modified concept definitions: 22 civil rights, election, voting system, intergovernmental elections, local elections, primary elections, referendum, regional elections, voting, fundamental rights, censorship and freedom of speech, freedom of religion, freedom of the press, human rights, football, political debates, privacy, women’s rights, breaking (breakdance)

1 hierarchy move: fundamental rights has been moved from politics to society.

Also, the Wikidata mapping URIs have all been changed to point to the http:// version of the URI instead of the https:// version. This follows the official Wikidata guidance.

See the official Media Topic vocabulary on the IPTC Controlled Vocabulary server, and an easier-to-navigate tree view. An Excel version of IPTC Media Topics is also available.

Digital Source Type updates

5 new concepts have been added:

• Multi-frame computational capture sampled from real life, intended to cover media recorded by modern cameras and smartphones that may process several captured images together to create the saved media file, without any interaction with the photographer. 
• Human-edited media, intended to replace the retired Original media with minor human edits, given that it is subjective to decide what is a “minor” edit.
• Digital creation, intended to replace the retired Digital art so that we can avoid the existential question of “what is art?”
• Screen capture, covering screenshots and screen recordings made on a device
• Composite of elements, as a generic form of the more specific “composite” terms.

2 concepts have been retired: Original media with minor human edits, and Digital art, as explained above.

8 concepts have had their names and definitions modified, while retaining the same machine-readable ID for backwards-compatibility purposes:

• Digital capture sampled from real life (ID: digitalCapture), replacing the previous name “Original digital capture sampled from real life”
• Digitised from a transparent negative (ID: negativeFilm), replacing the previous name “Digitised from a negative on film”
• Digitised from a transparent positive (ID: positiveFilm), replacing the previous name “Digitised from a positive on film”
• Digitised from a non-transparent medium (ID: print), replacing the previous name “Digitised from a print on non-transparent medium”
• Edited using Generative AI (ID: compositeWithTrainedAlgorithmicMedia), replacing the previous name “Composite with Trained algorithmic media”
• Algorithmically-altered media (ID: algorithmicallyEnhanced), replacing the previous name “Algorithmically Enhanced”
• Created using Generative AI (ID: trainedAlgorithmicMedia), replacing the previous name “Trained Algorithmic Media”
• Virtual event recording (ID: virtualRecording), replacing the previous name “Virtual recording”

Our thanks go to IPTC representatives and experts from Partnership on AI, Google, Adobe, C2PA, CIPA and many others on making these updates to our vocabulary, which is now widely used to identify Generative AI content.

Updates to other NewsCodes vocabularies

Alternative  Identifier Role (altidrole)

• Vocabulary’s name changed to fix a spelling mistake.
• New concept: IPTC Video Metadata Hub ID (altidrole:vmhVideoId)

Event Occur Status (eocstat)

• Fix spelling mistake “occurence” -> “occurrence” throughout.

Golf Shot (spgolshot)

• New concept: Chip (spgolshot:chip)

Rights Property (rightsprop)

• New concept: Copyright Year (rightsprop:copyrightyear)
• 4 modified definitions: Minor Model Age Disclosure, Model Release Id, Model Release Status, Property Release Status.

Sports Concept (spct)

• New concept: Recurring Competition (spct:recurring-competition)
• New concept: Governing Body (spct:governing-body)

The IPTC NewsCodes Working Group has released the latest update to IPTC NewsCodes vocabularies.

The changes are quite minor this time, but we still recommend that users stay up to date with the latest version.

Changes to Media Topics vocabulary

Our main subject classification taxonomy, IPTC Media Topics, has seen the following updates:

1 new concept

• breaking (breakdance) (added earlier this year in time for the Paris 2024 Olympics)

1 retired concept

• missing in action (duplicate term added in error in the 2024 Q1 update. The existing term missing in action medtop:20000061 was moved to replace the newer term))

32 modified definitions

These changes mostly correct spelling errors in en-GB where US spellings had slipped in, such as changing “behavior” to “behaviour” for en-GB:

wireless technology, tobacco and nicotine, economic trends and indicators, international economic institution, stocks and securities, adult and continuing education, upper secondary education, social learning, medical condition, Confucianism, relations between religion and government, road cycling, competitive dancing, sexual misconduct, developmental disorder, fraternal and community group, cyber warfare, public transport, taxi and ride-hailing, shared transport, business reporting and performance, business restructuring, commercial real estate, residential real estate, podcast, financial service, business service, news industry, diversity, equity and inclusion, sustainability, profit sharing, breaking (breakdance).

As usual, the Media Topics vocabularies can be viewed in the following ways:

• In a collapsible tree view
• As a downloadable Excel spreadsheet
• On one page on the cv.iptc.org server
• In machine readable formats such as RDF/XML and Turtle using the SKOS vocabulary format: see the cv.iptc.org guidelines document for more detail.

Updates to other vocabularies

Horse Position (sphorposition)

New term “trainer” added to https://cv.iptc.org/newscodes/sphorposition. This term is needed by IPTC Sport Schema.

For more information on IPTC NewsCodes in general, please see the IPTC NewsCodes Guidelines.

Last week, the IPTC Spring Meeting 2024 brought media industry experts together for three days in New York City to discuss many topics including AI, archives and authenticity.

Hosted by both The New York Times and Associated Press, over 50 attendees from 14 countries participated in person, with another 30+ delegates attending online.

As usual, the IPTC Working Group leads presented a summary of their most recent work, including a new release of NewsML-G2 (version 2.34, which will be released very soon); forthcoming work on ninjs to support events, planned news coverage and live streamed video; updates to NewsCodes vocabularies; more evangelism of IPTC Sport Schema; and further work on Video Metadata Hub, the IPTC Photo Metadata Standard and our emerging framework for a simple way to express common rights statements using RightsML.

We were very happy to hear many IPTC member organisations presenting at the Spring Meeting. We heard from:

• Anna Dickson of recently-joined member Google talked about their work with IPTC in the past and discussed areas where we could collaborate in the future
• Aimee Rinehart of Associated Press presented AP’s recent report on the use of generative AI in local news
• Scott Yates of JournalList gave an update on the trust.txt protocol
• Andreas Mauczka, Chief Digital Officer at Austria Press Agency APA presented on APA’s framework for use of generative AI in their newsroom
• Drew Wanczowski of  Progress Software gave a demonstration of how IPTC standards can be implemented in Progress’s tools such as Semaphore and MarkLogic
• Vincent Nibart and Geert Meulenbelt of new IPTC Startup Member Kairntech presented on their recent work with AFP on news categorisation using IPTC Media Topics and other vocabularies
• Mathieu Desoubeaux of IPTC Startup Member IMATAG presented their work, also with AFP, on watermarking images for tracking and metadata retrieval purposes

In addition we heard from guest speakers:

• Jim Duran of the Vanderbilt TV News Archive spoke about how they are using AI to catalog and tag their extensive archive of decades of broadcast news content
• John Levitt of Elvex spoke about their system which allows media organisations to present a common interface (web interface and developer API) to multiple generative AI models, including tracking, logging, cost monitoring, permissions and other governance features which are important to large organisations using AI models.
• Toshit Panigrahi, co-founder of TollBit spoke about their platform for “AI content licensing at scale”, allowing content owners to establish rules and monitoring around how their content should be licensed for both the training of AI models and for retrieval-augmented generation (RAG)-style on-demand content access by AI agents.
• We also heard an update about the TEMS – Trusted European Media Data Space project. 

We were also lucky enough to take tours of the Associated Press Corporate Archive on Tuesday and the New York Times archive on Wednesday. Valierie Komor of AP Corporate Archives and Jeff Roth of The New York Times Archival Library (known to staffers as “the morgue”) both gave fascinating insights and stories about how both archives preserve the legacy of these historically important news organisations.

Brendan Quinn, speaking for Judy Parnall of the BBC, also presented an update of the recent work of C2PA and Project Origin and introduced the new IPTC Media Provenance Committee, dedicated to bringing C2PA technology to the news and media industry.

On behalf all attendees, we would like to thank The New York Times and Associated Press for hosting us, and especially to thank Jennifer Parrucci of The New York Times and Heather Edwards of The Associated Press for their hard work in coordinating use of their venues for our meeting.

The next IPTC Member Meeting will be the 2024 Autumn Meeting, which will be held online from Monday September 30th to Wednesday October 2nd, and will include the 2024 IPTC Annual General Meeting. The Spring Meeting 2025 will be held in Western Europe at a location still to be determined.

The latest update to IPTC NewsCodes, the 2024-Q1 release, was published on Thursday 28th March.

This release includes many updates to our Media Topic subject vocabulary, plus changes to Content Production Party Role, Horse Position, Tournament Phase, Soccer Position, Genre, User Action Type and Why Present.

UPDATE on 11 April: we released a small update to the Media Topics, including Norwegian (no-NB and no-NN) translations of the newly added terms, thanks to Norwegian news agency NTB.

We also made one label change in German: medtop:20000257 from “Alternative-Energie” to “Erneuerbare Energie,” This change was made at the request of German news agency dpa.

Changes to Media Topics vocabulary

As part of the regular review undertaken by the NewsCodes Working Group, many changes were made to the economy, business and finance branch of Media Topics. In addition, a number of changes were made to the conflict, war and peace branch in response to suggestions made by new IPTC member ABC Australia.

5 new concepts: sustainability, profit sharing, corporate bond, war victims, missing in action.

12 retired concepts: justice, restructuring and recapitalisation, bonds, budgets and budgeting, consumers, consumer issue, credit and debt, economic indicator, government aid, investments, prices, soft commodities market.

55 modified concepts: peacekeeping force, genocide, disarmament, prisoners of war, war crime, judge, economy, economic trends and indicators, business enterprise, central bank, consumer confidence, currency, deflation, economic growth, gross domestic product, industrial production, inventories, productivity, economic organisation, emerging market, employment statistics, exporting, government debt, importing, inflation, interest rates, international economic institution, international trade, trade agreements, balance of trade, trade dispute, trade policy, monetary policy, mortgages, mutual funds, recession, tariff, market and exchange, commodities market, energy market, debt market, foreign exchange market, loan market, loans and lending, study of law, disabilities, mountaineering, sport shooting, sport organisation, recreational hiking and climbing, start-up and entrepreneurial business, sharing economy, small and medium enterprise, sports officiating, bmx freestyle.

48 concepts with modified names/labels: judge, emergency incident, transport incident, air and space incident, maritime incident, railway incident, road incident, restructuring and recapitalisation, economic trends and indicators, exporting, importing, interest rates, balance of trade, mortgages, commodities market, soft commodities market, loans and lending, study of law, disabilities, mountain climbing, mountaineering, sport shooting, sport organisation, recreational hiking and climbing, start-up and entrepreneurial business, sports officiating, bmx freestyle, tsunami, healthcare industry, developmental disorder, depression, anxiety and stress, public health, pregnancy and childbirth, fraternal and community group, cyber warfare, public transport, taxi and ride-hailing, shared transport, business reporting and performance business restructuring commercial real estate residential real estate podcast, financial service, business service, news industry, diversity, equity and inclusion. 

57 modified definitions: war crime, economy, economic trends and indicators, business enterprise, central bank, consumer confidence, currency, deflation, economic growth, economic organisation, emerging market, employment statistics, exporting, government debt, importing, inflation, interest rates, international economic institution, trade agreements, trade dispute, trade policy, mortgages, recession, tariff, market and exchange, commodities market, energy market, soft commodities market, debt market, foreign exchange market, loan market, loans and lending, disabilities, mountaineering, sport organisation, start-up and entrepreneurial business, sharing economy, small and medium enterprise, tsunami, healthcare industry, developmental disorder, depression, anxiety and stress, public health, pregnancy and childbirth, cyber warfare, public transport, taxi and ride-hailing, shared transport, business reporting and performance, business restructuring, commercial real estate, residential real estate, podcast, financial service, business service, news industry.

22 modified broader terms (hierarchy moves): peacekeeping force, genocide, disarmament, prisoners of war, business enterprise, central bank, consumer confidence, currency, gross domestic product, industrial production, inventories, productivity, economic organisation, emerging market, interest rates, international economic institution, international trade, monetary policy, mutual funds, tariff, loans and lending, bmx freestyle.

These changes are already available in the en-GB, en-US and Swedish (se) language variants. Thanks go to TT and Bonnier News for their work on the Swedish translation.

If you would like to contribute or update a translation to your language, please contact us.

Sports-related NewsCodes updates

We also made some changes to our sports NewsCodes vocabularies, which are mostly used by SportsML and IPTC Sport Schema.

New vocabulary: Horse Position

New entries in Tournament Phase vocabulary: Heat, Round of 16

New entry in Soccer Position: manager,

News-related NewsCodes updates

Content Production Party Role: new term Generative AI Prompt Writer which can also be used in Photo Metadata Contributor to declare who wrote the prompt that was used to generate an image.

Genre: new term User-Generated Content.

Why Present: new term associated.

The User Action Type vocabulary, mostly used by NewsML-G2, has had some major changes.

Previously this vocabulary defined terms related to specific social media services or interactions. We have retired/deprecated all site-specific terms (Facebook Likes, Google’s +1,  Twitter re-tweets, Twitter tweets).

Instead, we have defined some generic terms: Like, Share, Comment. The pageviews term has been broadened into simply views (although the ID remains as “pageviews” for backwards-compatibility)

Thanks to the NewsCodes Working Group for their work on this release, and to all members and non-members who have suggested changes.

As we wrap up 2023, we thought it would be useful to give an update you on the IPTC’s work in 2023, including updates to most of our standards.

Two successful member meetings, one in person!

This year we finally held our first IPTC Member meeting in person since 2019, in Tallinn Estonia. We had around 30 people attend in person and 50 attended online from over 40 organisations. Presentations and discussions ranged from the e-Estonia digital citizen experience to building re-usable news content widgets with Web Components, and of course included generative AI, credibility and fact checking, and more. Here’s our report on the  IPTC 2023 Spring Meeting.

For our Autumn Meeting we went back to an online format, with over 50 attendees, and more watching the recordings afterwards (which are available to all members). Along with discussions of generative AI and content licensing at this year’s meetings, it was great to hear the real-world implementation experience of the ASBU Cloud project from the Arab States Broadcasting Union. The system was created by IPTC members Broadcast Solutions, based on NewsML-G2. The DPP Live Production Exchange, led by new members Arqiva, will be another real-world implementation coming soon. We heard about the project’s first steps at the Autumn Meeting.

Also at this years Autumn Meeting we also heard from Will Kreth of the HAND Identity platform and saw a demo of IPTC Sport Schema from IPTC member Progress Software (previously MarkLogic). More on IPTC Sport Schema below! All news from the Autumn Meeting is summed up in our post AI, Video in the cloud, new standards and more: IPTC Autumn Meeting 2023

We’re very happy to say that the IPTC Spring Meeting 2024 will be held in New York from April 15 – 17. All IPTC member delegates are welcome to attend the meeting at no cost. If you are not a member but would like to present your work at the meeting, please get in touch using our Contact Us form.

IPTC Photo Metadata Conference, 7 May 2024: save the date!

Due to several issues, we were not able to run a Photo Metadata Conference in 2023, but we will be back with an online Photo Metadata Conference on 7th May 2024. Please mark the date in your calendar!

As usual, the event will be free and open for anyone to attend.

If you would like to present to the people most interested in photo metadata from around the world, please let us know!

Presentations at other conferences and work with other organisations

IPTC was represented at the CEPIC Congress in France, the EBU DataTech Seminar in Geneva, Sports Video Group Content Management Forum in New York and the DMLA’s International Digital Media Licensing Conference in San Francisco.

We also worked with CIPA, the organisation behind the Exif photo metadata standard, on aligning Exif with IPTC Photo Metadata, and supported them in their work towards Exif 3.0 which was announced in June.

The IPTC will be advising the TEMS project which is an EU-funded initiative to build a “media data space” for Europe, and possibly beyond: IPTC working with alliance to build a European Media Data Space.

IPTC’s work on Generative AI and media

Of course the big topic for media in 2023 has been Generative AI. We have been looking at this topic for several years, since it was known as “synthetic media” and back in 2022 we created a taxonomy of “digital source types” that can be used to describe various forms of machine-generated and machine-assisted content creation. This was a joint effort across our NewsCodes, Video Metadata and Photo Metadata Working Groups.

It turns out that this was very useful, and the IPTC Digital Source Type taxonomy has been adopted by Google, Midjourney, C2PA and others as a way to describe content. Here are some of our news posts from 2023 on this topic:

• IPTC publishes metadata guidance for AI-generated “synthetic media”
• Google announces use of IPTC metadata for generative AI images
• Midjourney and Shutterstock AI sign up to use of IPTC Digital Source Type to signal generated AI content
• Microsoft announces signalling of generative AI content using IPTC and C2PA metadata
• Royal Society/BBC workshop on Generative AI and content provenance
• New “digital source type” term added to support inpainting and outpainting in Generative AI
• IPTC releases technical guidance for creating and editing metadata, including DigitalSourceType

IPTC’s work on Trust and Credibility

After a lot of drafting work over several years, we released the Guidelines for Expressing Trust and Credibility signals in IPTC standards that shows how to embed trust infiormation in the form of “trust indicators” such as those from The Trust Project into content marked up using IPTC standards such as NewsML-G2 and ninjs. The guideline also discusses how media can be signed using C2PA specification.

We continue to work with C2PA on the underlying specification allowing signed metadata to be added to media content so that it becomes “tamper-evident”. However C2PA specification in its current form does not prescribe where the certificates used for signing should come from. To that end, we have been working with Microsoft, BBC, CBC / Radio Canada and The New York Times on the Steering Committee of Project Origin to create a trust ecosystem for the media industry. Stay tuned for more developments from Project Origin during 2024.

IPTC’s newest standard: IPTC Sport Schema

After years of work, the IPTC Sports Content Working Group released version 1.0 of IPTC Sport Schema. IPTC Sport Schema takes the experience of IPTC’s 10+ years of maintaining the XML-based SportsML standard and applies it to the world of the semantic web, knowledge graphs and linked data.

Paul Kelly, Lead of the IPTC Sports Content Working Group, presented IPTC Sport Schema to the world’s top sports media technologists: IPTC Sport Schema launched at Sports Video Group Content Management Forum.

Take a look at out dedicated site https://sportschema.org/ to see how it works, look at some demonstration data and try out a query engine to explore the data.

If you’re interested in using IPTC Sport Schema as the basis for sports data at your organisation, please let us know. We would be very happy to help you to get started.

Standard and Working Group updates

• Our IPTC NewsCodes vocabularies had two big updates, the NewsCodes 2023-Q1 update and the NewsCodes Q3 2023 update. For our main subject taxonomy Media Topics, over the year we added 12 new concepts, retired 73 under-used terms, and modified 158 terms to make their labels and/or descriptions easier to understand. We also added or updated vocabularies such as Digital Source Type and Authority Status.
• The News in JSON Working Group released ninjs 2.1 and ninjs 1.5  in parallel, so that people who cannot move from the 1.x schema can still get the benefits of new additions. The group is currently working on adding events and planning items to ninjs based on requirements the DPP Live Production Exchange project: expect to see something released in 2024.
• NewsML-G2 2.32 and NewsML-G2 v2.33 were released this year, including support for Generative AI via the Digital Source Type vocabulary.
• The IPTC Photo Metadata Standard 2023.1 allows rightsholders to express whether or not they are willing to allow their content to be indexed by search engines and data mining crawlers, and whether the content can be used as training data for Generative AI. This work was done in partnership with the PLUS Coalition. We also updated the IPTC Photo Metadata Mapping Guidelines to accommodate Exif 3.0.
• Through discussions and workshops at our Member Meetings in 2022 and 2023, we have been working on making RightsML easier to use and easier to understand. Stay tuned for more news on RightsML in 2024.
• Video Metadata Hub 1.5 adds the same properties to allow content to be excluded from generative AI training data sets. We have also updated the Video Metadata Hub Generator tool to generate C2PA-compliant metadata “assertions”.

New faces at IPTC

Ian Young of Alamy / PA Media Group stepped up to become the lead of the News in JSON Working Group, taking over from Johan Lindgren of TT who is winding down his duties but still contributes to the group.

We welcomed Bonnier News, Newsbridge, Arqiva, the Australian Broadcasting Corporation and Neuwo.ai as new IPTC members, plus a very well known name who will be joining at the start of 2024. We’re very happy to have you all as members!

We are always happy to work with more organisations in the media and related industries. If you would like to talk to us about joining IPTC, please complete our membership enquiry form.

Here’s to a great 2024!

Thanks to everyone who gave IPTC your support, and we look forward to working with you in the coming year.

If you have any questions or comments (and especially if you would like to speak at one of our events in 2024!), you can contact us via our contact form.

Best wishes,

Brendan Quinn
Managing Director, IPTC
and the IPTC Board of Directors: Dave Compton (LSE Group), Heather Edwards (The Associated Press), Paul Harman (Bloomberg LP), Gerald Innerwinkler (APA), Philippe Mougin (Agence France-Presse), Jennifer Parrucci (The New York Times), Robert Schmidt-Nia of DATAGROUP (Chair of the Board), Guowei Wu (Xinhua)

The IPTC News Codes Working Group has just released a new batch of changes to the IPTC NewsCodes family of controlled vocabularies.

Note that we skipped the Q2 update this year because there weren’t many changes, and also because there were already so many changes in Q1 of this year.

Media Topic changes

Here’s a summary of changes to Media Topic vocabulary:

• 2 new concepts: sustainability, profit sharing
• 3 retired concepts: justice, restructuring and recapitalisation, soft commodity
• 7 modified names (labels): restructuring and recapitalisation, soft commodity, study of law, sport shooting, sport organisation, recreational hiking and climbing, mountaineering, disabilities (German and Norwegian translations)
• 2 modified definitions: mountaineering, sport organisation

Change to Media Topic tree browser

We have made a small change to the Media Topic tree browser tool: we now display a small “i” icon next to the label name for terms that have notes defined.

The terms that have notes are usually retired terms, and the note gives the user information regarding which terms should be used instead of the retired term. But in other cases notes are used to help explain changes or clarify usage.

Changes to other vocabularies

Other vocabularies have also been updated:

• Content Production Party Role sees two new terms, contentEditor and metadataEditor, that can be used to show changes made by humans or systems (such as AI engines)
• Format had a small change to indicate that it is not just for NewsML 1 documents.
• User Action Type had a small bug fix, changed references to Twitter / X and retired Google Plus as a term. More changes will be coming soon covering other social media platforms and ways to track user interactions with media content.
• The rendition CV has been updated to make it more generic – renditions can apply to any type of media, not just images and video.
• The digitalsourcetype CV had already been updated in July to handle inpainting and outpainting but we mention it again here as a reminder.

Thanks to the representatives from IPTC members AFP, NTB, Bonnier News, ABC Australia, Bloomberg, New York Times and Associated Press for their contributions to the changes this quarter via the NewsCodes Working Group.

We are still working on our regular review of Media Topics – currently we are in the middle of a review of the Economy branch. The review is not yet complete but we hope for it to be ready for the Q4 or Q1 update.

The IPTC NewsCodes Working Group has approved an addition to the Digital Source Type NewsCodes vocabulary.

The new term, “Composite with Trained Algorithmic Media“, is intended to handle situations where the “synthetic composite” term is not specific enough, for example a composite that is specifically made using an AI engine’s “inpainting” or “outpainting” operations.

The full Digital Source Type vocabulary can be accessed from https://cv.iptc.org/newscodes/digitalsourcetype. It can be downloaded in NewsML-G2 (XML), SKOS (RDF/XML, Turtle or JSON-LD) to be integrated into content management and digital asset management systems.

The new term can be used immediately with any tool or standard that supports IPTC’s Digital Source Type vocabulary, including the C2PA specification, the IPTC Photo Metadata Standard and IPTC Video Metadata Hub.

Information on the new term will soon be added to IPTC’s Guidance on using Digital Source Type in the IPTC Photo Metadata User Guide.

Following the recent announcements of Google’s signalling of generative AI content and Midjourney and Shutterstock the day after, Microsoft has now announced that it will also be signalling the provenance of content created by Microsoft’s generative AI tools such as Bing Image Creator.

Microsoft’s efforts go one step beyond those of Google and Midjourney, because they are adding the image metadata in a way that can be verified using digital certificates. This means that not only is the signal added to the image metadata, but verifiable information is added on who added the metadata and when.

As TechCrunch puts it, “Using cryptographic methods, the capabilities, scheduled to roll out in the coming months, will mark and sign AI-generated content with metadata about the origin of the image or video.”

The system uses the specification created by the Coalition for Content Provenance and Authenticity. a joint project of Project Origin and the Content Authenticity Initiative.

The 1.3 version of the C2PA Specification specifies how a C2PA Action can be used to signal provenance of Generative AI content. This uses the IPTC DigitalSourceType vocabulary – the same vocabulary used by the Google and Midjourney implementations.

This follows IPTC’s guidance on how to use the DigitalSourceType property, published earlier this month.

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.