Overview of the C2PA trust ecosystem, showing how the C2PA project implements requirements set by both the Content Authenticity Initiative and Project Origin.
Overview of the C2PA trust ecosystem, showing how the C2PA project implements requirements set by both the Content Authenticity Initiative and Project Origin.

The IPTC is proud to announce that after intense work by most of its Working Groups, we have published version 1.0 of our guidelines document: Expressing Trust and Credibility Information in IPTC Standards.

The culmination of a large amount of work over the past several years across many of IPTC’s Working Groups, the document represents a guide for news providers as to how to express signals of trust known as “Trust Indicators” into their content.

Trust Indicators are ways that news organisations can signal to their readers and viewers that they should be considered as trustworthy publishers of news content. For example, one Trust Indicator is a news outlet’s corrections policy. If the news outlet provides (and follows) a clear guideline regarding when and how it updates its news content.

The IPTC guideline does not define these trust indicators: they were taken from existing work by other groups, mainly the Journalism Trust Initiative (an initiative from Reporters Sans Frontières / Reporters Without Borders) and The Trust Project (a non-profit founded by Sally Lehrman of UC Santa Cruz).

The first part of the guideline document shows how trust indicators created by these standards can be embedded into IPTC-formatted news content, using IPTC’s NewsML-G2 and ninjs standards which are both widely used for storing and distributing news content.

The second part of the IPTC guidelines document describes how cryptographically verifiable metadata can be added to media content. This metadata may express trust indicators but also more traditional metadata such as copyright, licensing, description and accessibility information. This can be achieved using the C2PA specification, which implements the requirements of the news industry via Project Origin and of the wider creative industry via the Content Authenticity Initiative. The IPTC guidelines show how both IPTC Photo Metadata and IPTC Video Metadata Hub metadata can be included in a cryptographically signed “assertion” 

We expect these guidelines to evolve as trust and credibility standards and specifications change, particularly in light of recent developments in signalling content created by generative AI engines. We welcome feedback and will be happy to make changes and clarifications based on recommendations.

The IPTC sends its thanks to all IPTC Working Groups that were involved in creating the guidelines, and to all organisations who created the trust indicators and the frameworks upon which this work is based.

Feedback can be shared using the IPTC Contact Us form.

Ian Young of PA Media / Alamy, lead of the IPTC News in JSON Working Group
Ian Young of PA Media Group / Alamy is the new Lead of the IPTC News in JSON Working Group.

The IPTC Standards Committee is happy to announce that ninjs, IPTC’s schema for marking up news content in JSON, has been revised to versions 2.1 and 1.5.

The vote to approve the new versions was taken at the recent IPTC Spring Meeting in Tallinn, Estonia and online.

This is in keeping with IPTC’s decision to maintain two parallel versions of ninjs: one for those who can’t upgrade to the 2.x version of backwards compatibility reasons, and those who prefer the simpler structure of ninjs 2.x that is easier to handle in some tools.

The ninjs User Guide has been updated to reflect the changes, which are summarised below.

ContactInfo added to ninjs 1.5 and 2.1

ninjs 2.1 and ninjs 1.5 both include the new contactinfo structure which can be used in the people, organisations, places and infosources properties (and their ninjs 1.x equivalents person, organisation, place and infosource).

The contactInfo structure can contain physical or online contact information such as a street address or postal address, a username on social media such as Twitter, Instagram or TikTok, or even a locator such as what3words.

Here are some examples of how the contactinfo property can be used:

"people": [
  {
    "name": "Jonas Svensson",
    "contactinfo": [
      {
        "type":"phone",
        "role": "work",
        "value": "+46 (0)8-7887500"
      }
    ]
  }
],

"organisations": [
  {
    "name": "International Committee of the Red Cross",
    "contactinfo": [
      {
        "type": "web",
        "value": "https://www.icrc.org/"
      },
      {
        "type": "address",
        "address": {
          "lines": [
            "19 Avenue de la paix",
            "1202 Geneva",
            "Switzerland"
          ]
        }
      },
      {
        "type": "telephone",
        "value": "+41 22 734 60 01"
      }
    ]
  }
]

Better support for organisation identifiers such as tickers, ISIN etc

ninjs 2.1 and 1.5 also include the new symboltype and symbol properties under symbols. Symbol can identify any type of URI describing the type of the symbol. The CV http://cv.iptc.org/newscodes/financialinstrumentsymboltype is recommended.

The ticker sub-property under symbols is now deprecated. This means that it can still be used if necessary, but use is not recommended.

We now recommend that ticker symbols are stored using symbol="TCKR" and symboltype="https://cv.iptc.org/newscodes/financialinstrumentsymboltype/Ticker".

Better support for machine classification

The subjects (ninjs 2.x) / subject (ninjs 1.x) properties now allow for the sub-properties creator, relevance and confidence

This allows organisations to more accurately use machine-generated subject tags in their content.  while stating that it was created by a machine (using the creator property), and giving numerical values for the relevance and confidence scores that are reported by machine tagging engines. (Of course, these properties can also be used for human-created subject tags if necessary!)

In addition, some internal changes to the schema were made to fix a validation bug that existed in previous versions. In order to accommodate these changes, the ninjs 2.1 schema uses the https://json-schema.org/draft/2020-12/schema version of JSON Schema.

Thanks to Johan Lindgren, welcome Ian Young as Working Group Lead

At the Spring Meeting in Tallinn we said farewell to Johan Lindgren as Lead of the News in JSON Working Group.

Johan, of the TT news agency in Sweden, was instrumental in bringing the News in JSON Working Group back from its quiet period after the initial launch of ninjs. This directly led to the release of several new versions of ninjs over the past few years, and its adoption by many of the world’s top news providers.

The IPTC wishes to thank Johan for all his contributions, and wishes him well for his retirement.

Johan’s work will be taken over by Ian Young from PA Media Group / Alamy based in the UK. Ian steps up to the Lead role after participating in the Working Group for many years, since the earliest days of ninjs.

We thank Ian for being willing to take on the lead role, and we look forward to seeing what developments will emerge from the News in JSON Working Group in the future.

Example of IPTC's ninjs format for syndicating news in JSON formatAt the IPTC Spring Meeting in May 2022, IPTC’s Standards Committee voted to approve ninjs 1.4, the latest version in the 1.x track of IPTC’s standard for news content in the JSON format.

Johan Lindgren of TT Nyhetsbyrån, Lead of the IPTC News in JSON Working Group, said:

“After the launch of ninjs 2.0 in the autumn of 2021, we received requests to add some of the new 2.0 features to the first generation of ninjs, so that those who are using the 1.x branch of ninjs can use the new features without making breaking changes. So we are excited to publish version 1.4 of ninjs, where these features are included.”

Those changes include:

  • New property contentcreated, denoting the date and time when the content of this ninjs object was originally created (as opposed to the date and time when the ninjs object itself was created). For example, an old photo that is now handled as a ninjs object may have a firstcreated and versioncreated of “2022-06-02T12:00:00+00:00”, but a contentcreated value of “1933-04-03T00:00:00+00:00”. The contents must be a valid JSON Schema date-time object.
  • New property expires, showing “the date and time after which the Item is no longer considered editorially relevant by its provider.” Note that this is not the same as a rights-related expiration, it simply conveys the desire of the content creator to highlight the content until a certain time. A good example might be a football match preview, which would no longer be editorially relevant after the game commences. The contents must be a valid JSON Schema date-time object.
  • New property rightsinfo, which holds an expression of rights to be applied to the content. It contains sub-properties langid (a URI which specifies the language used to specify rights such as RightsML or ODRL), and one of either linkedrights (containing a link to a remotely-hosted declaration of the rights associated with the content) or encodedrights (which includes an embedded encoding of the rights statements within the ninjs object).

Which version of ninjs should I choose for my project?

There might be some confusion since we have released ninjs 1.4 after the release of ninjs 2.0. Please note that this is simply an update to the 1.x branch of ninjs to make it easier for users who cannot upgrade to 2.x branch due to breaking changes.

If you are starting a new project that requires JSON-encoded news content, we recommend using ninjs 2.0. This version should be easiest for developers to work with.

If you are already using a 1.x version of ninjs, we recommend at least upgrading to version 1.4. This should be an easy change, because 1.4 is backwards-compatible with versions 1.0, 1.1, 1.2 and 1.3. We would also recommend upgrading to 2.0 if possible, but if not, 1.4 is the best version of the 1.x branch.

Supporting materials for ninjs 1.4 and ninjs 2.0 can be found at these locations:

Thanks to Johan and the IPTC News in JSON Working Group for working on this release.

Screenshot of a code editor showing an extract of a news story in ninjs 2.0 format

Today, IPTC announces the release of version 2.0 of the news industry’s standard for exchanging content in JSON: ninjs.

The new version introduces a completely new way of declaring multiple headlines, body texts and description fields, which is compatible with binary data serialisation formats such as Avro and Protocol Buffers.

“We are very excited about releasing the 2.0-version of News in JSON (ninjs),” says Johan Lindgren (TT), lead of the working group responsible for developing the standard. “When working on improving the 1.3 version, we realised that a number of suggestions would mean breaking changes and after some consideration we took that step. Now we have a version of ninjs that is better suited for APIs, databases like Elastic and conversion to binary methods like Protocol Buffers.”

The IPTC News in JSON Working Group has kept the original focus on two main use cases: data in transit and data at rest.

In recent years, more systems have started to convert from JSON formats into binary data serialisation protocols such as Avro and Protocol Buffers for data in transit. However ninjs 1.x couldn’t be converted into these protocols because of the dynamic way that keys could be defined, for example “headline_main” and “headline_subhead”. In ninjs 2.0, all properties are given well-defined names, so they can be converted into Protobufs schemas. The GitHub repository for ninjs now includes a demonstration of how ninjs 2.0 can be used with Protocol Buffers.

Other tools included in the repository are an example GraphQL server for ninjs and example XSLTs to convert from IPTC XML-based formats like NewsML and NITF.

The ninjs Generator tool has been updated to create ninjs 2.0. In fact, using the tool, users can switch between generating ninjs 1.3 and ninjs 2.0 output at the click of a radio button.

The official location of the ninjs 2.0 JSON Schema is https://iptc.org/std/ninjs/ninjs-schema_2.0.json.

A full list of the changes in ninjs 2.0 can be viewed in section 7.5 of the ninjs User Guide.

IPTC members and our guests have just finished a very busy 2021 edition of our IPTC Autumn Meeting. Held online over three days, the meeting was a mix of IPTC Working Group presentations, members presenting recent projects, and invited guest speakers on important topics in the news and media world.

Screenshot of Videre AI's video understanding and annotation system
Screenshot of Videre AI’s video understanding platform at IPTC Autumn Meeting

This year we heard member presentations from:

  • Honor Craig-Bennett of the BBC reporting on the Images Digital Asset Management system, based on the Guardian’s open-source GRID system. We heard from Andy Read about this system
  • Heather Edwards from Associated Press spoke about their project to replace their existing rules-based classification system
  • Mark Milstein from Microstocksolutions spoke about a new project he is working on to create “synthetic media” AI-generated images and videos based on textual descriptions and metadata
  • DATAGROUP Consulting Group’s Robert Schmidt-Nia spoke about a project using AWS’s Comprehend text classification service to power a serverless news classification system using IPTC’s Media Topics vocabulary
  • Frameright‘s Marina Ekroos speaking about an EU stars4media project they are working on called “Artificial Intelligence in photojournalism: can it work?”
  • Scott Yates from new Startup Member JournalList spoke about the trust.txt project, letting news providers state their affiliates and official social media channels in a simple way
  • Bruce MacCormack from CBC / Radio Canada spoke about Project Origin, looking at authenticity for video and news media, passing requirements to the C2PA work
  • The BBC‘s Charlie Halford spoke about C2PA, updating members with a deep technical view on how the system is planned to work, as detailed in the recently-released draft specification.

In addition, we heard from guest speakers:

  • Keesiu Wong of Design AI spoke about the Videre AI project, looking at “next-generation video understanding”. He was joined by project partner Javier Picazo from Associate Member Agencia EFE, Spain’s national news agency.
  • Alex Lakatos of Interledger spoke about the distributed payments technology which is used by…
  • Uchi Uchibeke of Coil who use Interledger to implement micropayments which can be implemented on publisher websites by adding one line of HTML.

New standard versions

The Working Group presentations were also packed with content, in particular three new standard versions that were proposed to the Standards Committee:

  • NewsML-G2 v2.30 adds fields for “residrefformat” and “residrefformaturi” to enable publishers to describe the format of a resource ID reference, and makes catalog and catalogRef optional to support publishers who only use URIs for controlled values and therefore have no need for catalogs
  • The News in JSON Working Group’s ninjs v2.0 is a non-backwards-compatible new release which changes the way repeating values are handled, moving from patternProperties fields with arbitrary names such as “body_text” and “body_html” to arrays with fixed names such as “bodies”. The objects within the array elements include properties “role” and “contenttype” which take the place of the arbitrary extension to the “body_” tag.
  • The IPTC Photo Metadata Standard v2021.1 adds new properties to IPTC Core which are intended to be used for accessibility purposes: “Alt Text (Accessibility)” and “Extended Description (Accessibility)”. We have also added and Event Identifier property to align with other metadata ID properties, and modified the Description Writer field to include the writer of the accessibility fields.

New faces

We were very happy to welcome new members Frameright, JournalList, Spotlight Sports Group, Glide Publishing Platform to the meeting.

The Standards Committee was chaired for the first time by new Chair Paul Harman of Bloomberg.

The AGM was the first for new Treasurer, Gerald Innerwinkler of Austria Press Agentur APA.

And we congratulate Philippe Mougin of Agence France-Presse AFP for being voted on to the IPTC Board of Directors, along with the existing Board members who were all re-elected.

It was another great meeting with over 70 representatives from 42 organisations in 17 different countries! We’re hoping that the next IPTC member meeting will be back to face-to-face, and we have provisionally booked Tallinn, Estonia for 16 – 18 May, 2022. We will confirm this in January 2022.

Example of IPTC's ninjs format for syndicating news in JSON formatBased on discussions at the recent IPTC Autumn Meeting, the IPTC News in JSON Working Group is updating its view of the use of ninjs and other forms of JSON for handling news content.

If your organisation uses JSON in any way for handling news content, we would like to hear from you.

We are looking for input from IPTC members and non-members, from agencies, publishers, broadcasters and software vendors.

Please help us by filling in the short survey via this Google Form.

The ninjs generator tool has been updated to the new ninjs 1.3

At last week’s Spring Meeting IPTC updated the News in JSON standard (ninjs) to version 1.3. The JSON schema of the new version can be accessed at https://www.iptc.org/std/ninjs/.

The updated schema now has support for trust indicators, genre, other types of headlines and a way for providers to enter their own alternative IDs.

Version 1.3 is backwards-compatible with previous versions of ninjs and makes no breaking changes.

It includes the following new properties and structures:

  • genre follows the structure of other objects in ninjs with the possibility to add a code, a name and a reference to a scheme of the code to indicate the genre of the news item.
  • trustindicator is also an object with properties to indicate and point at documents describing the providers status according to defined trust indicators. Read more about trust indicators here.
  • There has been a demand for other types of headlines, such as “subhead” or “mobile headline”. The original headline property is still in the schema. But in version 1.3 it is now accompanied by a headline_ construct which works in the same manner as body_ and description_. Providers can now add other types of headlines and name them to indicate format and/or type.
  • altid is a property that is open to the provider’s own definition of both names and types of sub-properties. That way providers can include alternative IDs as they originally appeared.
  • The 1.3 schema also includes a $standard object which contains properties for name, version and reference to the schema that the item follows. This is not (yet) supported in software tools but the idea is that tools could look up the schema for which a document is written, similarly to the way that XML allows users to state the XML Schema that should validate a document.

The user guide, sample generator and GraphQL sample have all been updated according to the 1.3 additions.

ninjs 1.3 will soon be included in the SchemaStore.org JSON Schema repository, to aid with editing and validation of ninjs 1.3 files in a range of popular code editors such as Visual Studio Code and Visual Studio 2013+, IntelliJ IDEA, PyCharm and PHPStorm.

For more information, please see:

If you have any questions or comments, please contact the News in JSON Working Group via the public ninjs discussion group, or contact IPTC via the Contact Us form.

The draft guidelines document

At the IPTC Autumn Meeting in Toronto in 2018, IPTC considered the issues of “trust and credibility” in news media. We looked at the existing initiatives and considered whether IPTC could contribute to the space.

We concluded that some existing efforts were doing great work and that we should not create our own trust and credibility standard. Instead, our resources could best be put towards working with those groups, and aligning IPTC’s standards — particularly our main news standards NewsML-G2 and ninjs —  to work well with the outputs of those groups.

Since that time, the IPTC NewsML-G2 Working Group has been collaborating with several initiatives around trust and misinformation in the news industry. We have been working mainly with The Trust Project and the Journalism Trust Initiative from Reporters Without Borders, but have also been in communication with the Credibility Coalition, the Certified Content Coalition and others to identify all known means of expressing trust in news content.

Our aim is to make it easy for users of NewsML-G2 and ninjs to work with these standards to convey the trustworthiness of their content. This should make it easier for news publishers to translate trust information to something that can be read by aggregator platforms and user tools.

In particular, we want to make it as easy as possible for syndicated content to be distributed and published in alignment with trust principles.

A new IPTC Guideline document

To that end, we are publishing a “public draft” of a new IPTC guideline document: Expressing Trust and Credibility Information in IPTC Standards. While not complete, we hope that it helps IPTC members and other users of our standards to understand how they can express trust indicators.

To go along with the draft, we are proposing some changes to existing IPTC standards, including updates to NewsML-G2 and to ninjs, and a new Trust Indicator taxonomy created as part of the IPTC NewsCodes.

New Genres in NewsCodes and changes to NewsML-G2 and ninjs

To accommodate the new work, we will be adding some new entries to the NewsCodes Genre vocabulary. Some genres required for this work such as “Opinion” and “Special Report” were already in the genres vocabulary, but we are proposing to add new genres including “Fact Check” and “Satire“, and some genres to handle sponsored content: Advertiser Supplied, Sponsored and Supported.

We will also be making some small changes to the existing ninjs and NewsML-G2 standards to accommodate some new requirements, such as being able to associate a publisher with another organisation, to indicate membership of The Trust Project, Journalism Trust Initiative or a similar group.

From trusted agency to publisher and then to a user

By following the guidelines, a news agency can add their own trust information to the news items that they distribute. A publisher can then take those trust indicators and convert them to the standard schema.org markup used to convey trust indicators in HTML pages (initially created via a collaboration between schema.org and The Trust Project in 2017).

The schema.org markup can then be read by search engines, platforms such as Facebook, and specialised trust tools such as the NewsGuard browser plugin, so that users can see the trust indicators and decide for themselves whether they can trust a piece of news.

Please give us your feedback

The document will not be final until after those changes have been approved by IPTC members at our next meeting in May.

We have published the draft to ask for feedback from the community about how we could improve our guidance, ask for any trust indicators that we have missed, and to ask for implementation feedback.

Please use the IPTC Contact Us form to send your feedback.

About the Trust Project

The Trust Project is a global network of news organizations working to affirm and amplify journalism’s commitment to transparency, accuracy and inclusion. The project created the Trust Indicators, which are a collaborative, journalism-generated standard for news that helps both regular people and the technology companies’ machines easily assess the authority and integrity of news. The Trust Indicators are based in robust user-centered design research and respond to public needs and wants.

For more information, visit thetrustproject.org.

The Trust Project is funded by Craig Newmark Philanthropies, Democracy Fund, Facebook, Google and the John S. and James L. Knight Foundation.

About the Journalism Trust Initiative

The Journalism Trust Initiative aims at a healthier information space. It is developing indicators for trustworthiness of journalism and thus, promote and reward compliance with professional norms and ethics. JTI is led by Reporters Without Borders (RSF) in partnership with the European Broadcasting Union (EBU), the Global Editors Network (GEN) and Agence France Presse (AFP).

For more, visit https://jti-rsf.org/en/ 

 

Following many change requests submitted by news organisations all over the world, the IPTC News in JSON Working Group is happy to announce the 1.2 version of its standard ninjs.

ninjs generator
The new ninjs generator allows you to create your own ninjs documents by filling in a web form.

The JSON Schema of the new version can be accessed at https://www.iptc.org/std/ninjs/.

We also created a ninjs User Guide that will enable new and existing users to understand how to put ninjs to use at their organisation.

Version 1.2 is backwards-compatible with version 1.0 and 1.1 and makes no breaking changes.

It includes the following new properties and structures:

  • The firstcreated property goes along with the versioncreated property to specify the date/time when the first version of a news item was published.
  • The charcount and wordcount properties allow the publisher to specify the total character count and word count of a news item (excluding figure captions and metadata).
  • The slugline, property allows the publisher to specify a “slug”, a human-readable identifier for the item. (note that no conditions are placed on the usage of this property, usage is up to each publisher).
  • The ednote, property allows publishers to specify instructions to editors.
  • The infosource structure can specify one or many information sources for the news item. It is a metadata structure that can handle literal strings or values from a controlled vocabulary.
  • The title property handles “A short natural-language name for the item.”

Also the following sub-properties were added:

  • The value component was added to the allowed values for type to specify parts of a larger news item.
  • The description of the renditions property was changed to allow for any type of rendition, not just images, and new sub-properties duration and format were added to enable audio and video renditions (for example, an audio version of a text story).

We have also included a test and validation suite so new versions of the JSON Schema can automatically be checked for compliance and backwards compatibility.

ninjs 1.2 is now included in the SchemaStore.org JSON Schema repository, to aid with editing and validation of ninjs 1.2 files in a range of popular code editors such as Visual Studio Code and Visual Studio 2013+, IntelliJ IDEA, PyCharm and PHPStorm.

For more information, please see:

If you have any questions or comments, please contact the News in JSON Working Group via the public ninjs discussion group, or contact IPTC via the Contact Us form.