Events

Experts often choose between numbers and language when communicating estimates. We find that incentives to persuade increase language use by 25–29 percentage points. Experts slant language more than numbers, and receivers are effectively persuaded. Our findings suggest experts leverage language's imprecision to communicate slanted messages, with implications for expert communication in policy and public discourse.

Large Language Models produce false statements, and users encounter AI errors every day. These errors must be taken into account when using AI. It is thereby important to distinguish between errors, fake news, structural bias, and discrimination. This lecture focuses on errors.

"Preserving meaning" defines an essential operation of culture. It gathers time into form, it extends memory into transmission, it shapes continuity through vigilance.

In a context where generative artificial intelligence, deepfakes and cross-border influence operations are reshaping online information, it is essential to understand the emerging risks. In this webinar, experts from civil society, media, research and and foresight will discuss emerging challenges and possible European responses.

Elisa Ricci, ELLIS Fellow, FBK and UniTN will be among the speakers at FoMo 2026!

Co-located with the 11th IEEE European Symposium on Security and Privacy (EuroS&P 2026), the workshop aims at exploring security and risk assessment in identity management systems