AI to fight disinformation (RIA)

The Media Action Plan is developed around three areas: recovery, transformation, and enabling and empowerment. Research and innovation is key to the area of the transformation. Innovative digital solutions play an important role in ensuring a pluralistic access to trustworthy and meaningful information and quality content.

The consumption of news media has substantially increased during the COVID19 crisis, with people seeking out relevant and factual information in a time of uncertainty. Unfortunately, online services have also been used by malicious actors to spread dangerous disinformation campaigns, with sometimes fatal consequences. Similarly the need for reliable and consistent social media interactions as well as for access to audiovisual content, gaming and other interactive activities has increased during COVID-19.

Notably, media and digital literacy is one of the key pillars in fight against disinformation, as also identified e.g. in the Communication: “Tackling disinformation - the European Approach”. With the modern social media being more and more AI based, and contemporary disinformation mechanisms increasingly sophisticated, advanced means are required to ensure a trustworthy environment. AI technologies applied to tools and services tailored to the media ecosystem will help the access to and creation and distribution of trustworthy information and facilitate countering sophisticated manifestations of disinformation.

The outcomes from this topic will offer exploitation and take-up opportunities for the Digital Europe and Creative Europe Programmes.

Research and innovation proposals are expected to respond to one of the following:

1. Advanced AI based solutions for securing a trustworthy online environment. Disinformation techniques are already today strongly AI based. Therefore, scientific researchers and media practitioners need to be equipped with quantitative and semi-supervised tools based on AI, and network science driven tools of least same level of sophistication, capable of detecting different forms of deep-fakes and tampered content and to understand how and where such type of content spreads online. The development of such tools require the involvement of a scientific community at the intersection between AI and computer science, mathematics, social network sciences, social sciences and other relevant scientific fields, closely collaborating with journalists and media practitioners, and equipped with the necessary computing power to analyse rich content (e.g. videos and images) and to automate the management, processing and analysis of the flow of information within online systems.
2. Advanced AI based solutions targeted to citizens for securing a trustworthy online environment. The solutions should foster citizens’ ability to identify, verify and combat disinformation through AI innovation. Solutions provided would include the analysis and tracing of various forms of content, correlation/comparison of various sources of information, exploitation of contextual information. Efficient and intuitive mechanisms to convey the information regarding quality/veracity of information should be addressed, as appropriate. Transparency and human oversight would be key, with a view of empowering citizens.

Proposals should include, as appropriate, the development of tools and processes for design, testing and validation, deployment and uptake, auditing, certification (where appropriate), software engineering methodologies, as well as approaches to modularity and interoperability. Relevance to real-world applications should be demonstrated. Various approaches to AI for detection, elaboration of confidence levels, contradiction trade-offs, pattern identification in a field of heterogeneous sources and media formats, and also for decision support need to be considered. Involvement of multidisciplinary teams and transdisciplinary research will be essential. The consortia are strongly encouraged to team up with European companies, which are part of the media ecosystem, including SMEs, and also with non-media industrial and technological expertise.

Proposals should clearly identify the expected outcome it will focus on (i.e. targeting media professionals or citizens).

Proposals should also coordinate and ensure complementarity with the ongoing media and social media R&I, related to projects in the field (e.g. AI4Media[[https://cordis.europa.eu/project/id/951911]], Fandango[[https://cordis.europa.eu/project/id/780355]]) and take into consideration the expectations of the Media Action Plan and the European Democracy Action Plan, and collaborate with the European Digital Media Observatory. The media data space (developed under Digital Europe) infrastructures and communities might provide an opportunity to pilot new tools produced by the selected proposals. Proposals should foresee activities to collaborate with projects stemming from the topics HORIZON-CL3-2021-FCT-01-03: Disinformation and fake news are combated and trust in the digital world is raised.

All proposals are expected to embed mechanisms to assess and demonstrate progress (with qualitative and quantitative KPIs, demonstrators, benchmarking and progress monitoring), and share communicable results with the European R&D community, through the AI-on-demand platform, a public community resource, to maximise re-use of results and efficiency of funding.

Activities are expected to achieve TRL5-6 by the end of the project

All proposals are expected to allocate tasks to cohesion activities with the PPP on AI, Data and Robotics and funded actions related to this partnership, including the CSA HORIZON-CL4-2021-HUMAN-01-02.