News
/
July 2, 2025
Download PDF

Combating Online Misinformation: Launch of "Provenance for Trust"

The 1st collective initiative to guarantee the authenticity and traceability of media content, in support of credible information and renewed trust.

Paris, June 19, 2025 - Society and the media are going through an unprecedented crisis of confidence. Indeed, 48% of French people question the reliability of information relayed by the media. This growing mistrust is fuelled in particular by the proliferation of online disinformation and the rise of AI, making it difficult to discern the veracity of published content. And for good reason: despite technological advances in the sector, detection tools remain imperfect, automated moderation lacks nuance, and today's infobesity generates a volume of content that exceeds processing and verification capacities for content provenance and authenticity. Faced with these multiple challenges, a group of players including TrustMyContent, UncovAI, Journalism Trust Initiative (launched by Reporters sans Frontières), CEPIC, l'Atelier and the SciencePo Paris médialab today announced the launch of "Provenance For Trust", the 1st innovation program for the press and media committed to guaranteeing the authenticity, traceability and credibility of information.

"Provenance for Trust": combating misinformation and restoring trust

TrustMyContent, UncovAI, Journalism Trust Initiative (launched by Reporters sans Frontières), CEPIC, l'Atelier and SciencePo's médialab joined forces in 2024 to define the outlines of the"Provenance for trust" program.

At the end of this 1styear of work designed to understand and better grasp the challenges facing the sector, the collective addresses 6 concrete issues:

  1. AUTHENTICITY: How to verify the authenticity and source of content? How to build a genuine end-to-end chain of trust?
  2. COPYRIGHT: How can we guarantee and promote the origin of published content, and protect copyright?
  3. CERTIFICATION / LABELISATION: How can we highlight the transparency and quality of our processes and recognised certifications such as RSF's Journalism Trust Initiative?
  4. AI TRANSPARENCY: How can we transparently display the use of AI to our audience, and guarantee compliance with the AI Act?
  5. TYPOSQUATTING: How can we combat identity theft, websites and authentic content?
  6. ANTISCRAPING: How can we prevent the "scraping" of content from an official site and its transfer to AI-generated sites?

"It has become absolutely imperative to focus on the provenance and authenticity of content! Why? Because by ensuring a verifiable origin for every piece of content, users can more easily assess the reliability of the information they consume! This preventive approach eases the pressure on detection and moderation systems, fostering from the outset an environment where quality information is favoured. Working on provenance also strengthens public trust in media and institutions by clearly distinguishing authentic content from false information. While it doesn't solve all the problems, it does establish a more solid basis for the credibility of information, and that's the approach we're defending!"

- Mathieu Kervenec, spokesman for Provenance ForTrust

From tech to impact, the collective relies on its complementarity to build effective and sustainable solutions. Provenance for Trust will provide institutions and the media with a"toolbox" of technical solutions, methodologies and skills. To achieve this, the collective relies on the intervention and expertise of each of its current members:

  • TrustMyContent provides the technological foundation and the content labeling solution
  • UncovAI provides the AI-generated content detection solution
  • L'Atelier provides its development skills
  • Journalism Trust Initiative (JTI), an ISO-type standard enabling media to have the reliability of their editorial processes certified, provides the project with a catalogue of over a hundred certified media, and 700 media that have published their transparency report worldwide.

To measure the effectiveness of this new approach, SciencePo's médialab, also a partner in the project, is launching an impact study focusing on whether or not audiences regain confidence in authenticated and contextualised content.

Finally, CEPIC (International Association of the Media Licensing Industry), drawing on texts such as the Digital Services Act (DSA), the AI Act and the European Media Freedom Act (EMFA), as part of its work to research and promote the harmonization of practices and protection of the rights of creators of visual content worldwide, will address two major questions:

  • What regulatory mechanisms establish mandatory labelling requirements for AI-generated content?
  • Which regulatory frameworks suggest the need to label AI-generated content, even when this is not strictly mandatory?

Supported by the French Ministry of Culture and the FSEIP jury (Fonds de soutien à l'émergenceet à l'innovation dans la presse), Provenance For Trust is receiving funding to accelerate the program's development. So, for the next stages, the collective will in particular:

  • Launch, this summer 2025, a consultation of industry players, institutions and partners wishing to reflect on the use cases to be investigated and thus contribute to the program.
  • Propose that interested and relevant players join the Provenance for Trust collective.
  • Work in the medium term to build responses and set up appropriate "toolboxes" for institutions and the media.
  • Measure the impact and share the results with the Provenance for Trust community.

The program began in January 2025 and ends on December 31, 2026. This period should enable us to achieve our objectives of restoring lost confidence and providing concrete solutions.

Press Contact: contact@provenance4trust.org

A downloadable copy of this press release, which includes more information on each of these companies, is attached to this blog post: Provenance4Trust, TrustMyContent, L'ATELIER, UncovAI, Reporter Sans Frontière / Journalism Trust Initiative, CEPIC, and médialab de SciencesPo Paris.

More about CEPIC