During Switzerland’s presidency of the UN Security Council in October, we are organising with EPFL, the Swiss Federal Institute of Technology in Lausanne, and the International Committee of the Red Cross (ICRC) an exhibition called “Deepfake and you”.
Building sustainable peace and the protection of civilians in armed conflicts are priorities for Switzerland as an elected member of the Security Council for 2023-2024. To achieve these objectives, we believe it is crucial to take decisive action against misinformation, disinformation and hate speech. Through this exhibition, we aim to raise awareness about these risks, with a focus on the deepfake technology.
The exhibition is taking place at the Delegates' entrance of the UN Headquarters (accessible with a UN badge) from October 15 to 25.
Projects and possible solutions
The Swiss Confederation, the International Committee of the Red Cross and the EPFL are working with various partners to develop responses to threats posed by image manipulation. Some examples can be found here.
Deepfakes in Armed Conflict: Gender Perspectives
Disinformation and deepfakes disproportionally affect women and marginalized groups. The aim of this ongoing project is to explore the gendered dimension of these security threats in contexts of armed conflict. Stakeholders from various knowledge communities and across sectors will work together to explore how new technologies can be used to mitigate the risks of disinformation enabled by AI.
Stop Hate Speech
Free speech and open debate are core values of a democracy and are threatened by hate speech, notably in the comments section of social media and online news platforms, creating an environment where people do not feel safe to express themselves freely. The project aims to address this issue with an algorithm developed to detect hate speech online. It also encourages civil society to respond with counter-speech to curb hatred and make online debate more accessible to all, thereby strengthening freedom of expression.
ICAIN: Rebalancing AI Landscape
The goal of this global network is to democratize access to the world's leading supercomputing resources to develop AI models that benefit society worldwide and support the achievement of the UN Sustainable Development Goals.
Pilot projects are currently underway, including:
- Improving weather forecasting through AI for sustainable agriculture with Data Science Africa;
- Developing an ethical AI tool for humanitarian action with the ICRC, to ensure a reliable source tailored to its mandate.
Detecting and Countering Image Manipulation
Several projects and studies are underway to better understand the threat posed by the combination of cyber influence operations and AI, and to develop countermeasures to protect against these threats. For example, a project to create a catalogue of visually perceptible artifacts (i.e. defects in the image, caused by the AI generation process) that will make it possible to distinguish real photos from AI-generated images. In the future, the catalogue could be integrated into a workflow to help analysts verify the authenticity of image data.
