Workshop Trustworthy AI Algorithms for Media Ethics and Accountability

30th April 2024, 9:00h - 14:00h. Free registration (required).

Room 51.100 (La Nau). UPF Poble Nou Campus (Roc Boronat 138, Barcelona) and online.

Chaired by Vladimir Estivill-Castro and Carlos Castillo (DTIC-UPF), in collaboration with Ruth Rodríguez Martinez and Marcel Mauri (Dept. Communication, UPF). Guest speakers: Diego Saenz Trumper (Wikipedia), Nestor Duch-Brown (EC JRC), Coral García (Maldita.es), Mateusz Labuz (Ministry of Foreign Affairs, Republic of Poland), Lucas Gave (University of Wisconsin–Madison).

This workshop will allow experts in artificial intelligence and media to debate their research on the authenticity and trustworthiness of the creative outcomes when AI technologies are used in producing news, documents, research articles, or any other media or information source that is expected to be informative. The presenters will also express potential new challenges and research topics that can configure a medium and long term agenda of innovation for achieving trust in communication of news and information even if assisted with artificial intelligence.

Generative AI has revolutionised the capacity to collaboratively generate content. Teams of human authors using sophisticated prompts are creating media. While many times this could be explicit fantasy for entertainment; however, issues emerge when such creative process involves artefacts (news, summaries, discovery, citation finding, and even scientific reports or articles) where those consuming these artefacts expect them to be accurate, realistic, reliable and trustworthy. Thus, many research questions emerge, for instance, can the AI be generative but aligned with some particular moral values? Can those involved (humans and tools) be taught not to exhibit biases that obscure the truth? Can the AI itself be applied for fact checking? Since the production of content that ought to be trustworthy involves professionals in media and communication and the technology is evolving rapidly, the objective of the workshop is to create a forum for discussion between experts actively concerned for the issues that are already impacting the practices of today while considering even future challenges and potential solutions.

Talks will be in English and Spanish.

9:00-09:15 Welcome, by CESANI chair: Anders Johnson / and ELLIS Unit BCN Representative Gergo Neu

09:15:09.30 Welcome-Introduction by Workshop Organisers, 

Mini coffee/drinks break

09:45-10:15 Wikipedia:Verifiability, not truth 

Diego Saez Trumper, Senior Research Scientist at Wikimedia Foundation,  Wikipedia

10:15-10:45 AI to fight bad uses of AI, such as "fakes", 

Nestor Duch-Brown, Scientific Officer at Joint Research Centre of the European Commission 

10:45:11:15 Challenges that "fact checkers" face with the rise of AI and other technologies, 

Coral Garcia, International Alliances at Maldita.es

11:15-11:45 Regulating deepfakes in the Artificial Intelligence Act

Mateusz Łabuz, Ministry of Foreign Affairs of the Republic of Poland

30 min coffee break

12:15 -12:45 AI and the future of fact-checking

Lucas Graves, Professor at University of Wisconsin–Madison

12.45 -13:15  Disinformation, AI and Media Ethics Challenges

Ruth Rodríguez, Marcel Mauri (Dept. Communication, UPF)

13.15 -13:45 Value alignment for Machine Ethics (machine and human in the loop)

Vladimir Estivill-Castro

13.45-14.00 Final conclusions

Sponsored by ELLIS UNIT Barcelona, DTIC-UPF María de Maeztu Strategic Research Program, and the Center for Studies UPF in Artificial and Natural Intelligence (CeSANI).