Sections

Generative AI: what to believe and how to choose

SOLARIS project studying AI dangers and opportunities kicks off

Redazione Ansa

By Alessio Jacona* (ANSA) - ROME, APR 19 - The European SOLARIS project to study how to manage the threats and opportunities that generative artificial intelligence present to democracy, political engagement and digital citizenship has kicked off, featuring ANSA among the partners.
    "The risk is not only that of ending up believing things that are not true, but also, and above all, of not being able to believe in anything anymore," said SOLARIS project Federica Russo.
    This sums up perhaps the most serious threat posed by infodemics - the uncontrolled circulation of large quantities of unverified and often false information - when it is enhanced by generative artificial intelligence: making true and false almost (if not entirely) indistinguishable.
    SOLARIS is a European research project, which over the next three years aims to define methods and strategies to manage the risks and threats, but also the opportunities, that generative artificial intelligence present to democracy, political engagement and digital citizenship.
    Conceived exactly one year ago with a focus on generative adversarial networks (GANs) and their ability to generate increasingly realistic deepfake images and videos, after winning a European call for proposals, Solaris is now expanding its focus to other generative AI technologies, such as Midjourney and ChatGPT.
    "In the last few months we have witnessed an explosion of the generative AI phenomenon," said Russo, who is a philosopher of science, technology and information with the University of Amsterdam and University College London, "a sort of democratisation that has drastically lowered the access threshold" effectively multiplying the tools that can be used to generate fake-but-credible content.
    "Among the objectives of SOLARIS is to understand why some contents are more credible than others, regardless of the tools used to produce them," she continued.
    "Our studies already tell us that it is not only a question of image quality, but it is about the system too.
    "Which platform the video is shared on and where and who it comes from also makes the difference, as well as the characteristics of the user".
    An ecosystem problem.
    In short, thoroughly understanding the dynamics and functioning of a digital ecosystem in which a deepfake is successful is crucial to countering the phenomenon. In order to succeed "it is necessary to analyse not only the artefact, but also the entire system that the artefact is part of: a systemic approach, with which we aim to put users back at the centre, to eventually make them capable of evaluating the content they use for themselves, to distinguish the true content from the false". And, therefore, to be able to create a model of 'safe' use of content that can also be exported to other situations.
    "I would like to make it clear that, with our work, we do not intend to demonise technology per se,' Federica Russo continued to explain.
    "Instead we want to affirm the principle that it must be governed, not endured. My personal position is that technology must also be planned in the short and medium term, which is why it is so urgent to work on the training of developers, as well as on regulation'.
    The SOLARIS project partners.
    ANSA has the job of documenting and disseminating the activities and progress of the SOLARIS project over its three years of activity, which it will do (starting with this article) with news and in-depth reports published on the pages of the Artificial Intelligence Observatory. The other members of the project are the University of Amsterdam (which is leading the project), the National Interuniversity Consortium for Informatics (CINI), the University of Maribor in Slovenia, the Libera Università LUMSA, the University of Exeter in the UK, the Italian human-centred AI company DEXAI, the Carlos III University in Spain, the Albanian Institute for International Studies, the publisher Brand Media Bulgaria, the Albanian Ministry of the Interior and the European Citizen Science Association.
    What are GANs? Generative Adversarial Networks (GANs) - the main focus of the SOLARIS project's research - are a class of artificial intelligence model capable of creating multimedia content - audio and video - similar to real ones. Although there are several promising application areas for GANs, from audio-graphic productions to artistic creative expression, their current and planned misleading uses are equally numerous and worrying. This is the case with so-called 'deepfakes', fake images or videos that simulate real events with extreme precision. If trained on a face, for example, a GAN can make it move and speak in a hyper-realistic manner, and so it is not difficult to imagine why this technology could be used (and perhaps already has been) to spread fake news and disinformation. Hence the urgent need to develop practices that enable governments and legislators to improve the trustworthiness, transparency and reliability of GANs.
    The three main objectives of SOLARIS.
    SOLARIS takes up the challenge by giving itself three main objectives: to gain a clear understanding of how powerful GANs and other generative AI are from a technical point of view and under what conditions the content generated by these systems is perceived as trustworthy; to define regulatory innovations and policy options to address the political risks arising from the deployment of these technologies; to co-design with citizen science 'Generative AI for good', i.e. value-based systems to improve democratic engagement and digital citizenship.
    *Journalist, innovation expert and editor of the ANSA.it Artificial Intelligence Observatory (ANSA).
   

Leggi l'articolo completo su ANSA.it