Research fellows
Dirpolis Institute
Irina Carnat
Bio
In conjunction with the National PhD in Artificial Intelligence for Society, Irina joined the LIDER-Lab at Sant'Anna School of Advanced Studies in 2022, researching on accountability frameworks for the development of trustworthy AI systems.
Her research interests include AI governance from a comparative perspective, product liability for AI-embedded products, algorithmic accountability, and the impact of Generative AI on fundamental rights. She also collaborates within the Predictive Justice project, supervising the annotation process of legal data and developing solutions based on generative Large Language Models for the legal domain. As an intern in SmartLex, a university spinoff from Sant'Anna School of Advanced Studies, she partakes in designing, developing, testing, and validating an automatic anonymization tool using NER technologies, and gaining experience in product compliance for AI systems from design to market placement. During her university and doctoral studies, she has been a visiting fellow at the Institute for Information Law (IViR) and the Amsterdam Center for Transformative Private Law (ACT) of the University of Amsterdam, as well as at the University of Liège for the Erasmus exchange program.
Irina is a polyglot and licensed to practice law in Italy.
Publications
Carnat I., Comandé G., Licari D., De Nigris C., 'From human-in-the-loop to LLM-in-the-loop for high-quality legal dataset', 2024, in i-lex - Rivista di Scienze Giuridiche, Scienze Cognitive e Intelligenza Artificiale, Vol. 17, n. 1.
Carnat I., 'Intelligenza Artificiale e Responsabilità Civile', in Claudio Scognamiglio (ed.), Responsabilità Civile, Enciclopedia del Diritto (the Italian Encyclopedia of Law), November 2024.
Carnat I., 'Addressing the Risks of Generative AI for the Judiciary: the Accountability Framework(s) under the EU AI Act', 2024, in Computer Law & Security Review, vol. 55, November 2024.
Carnat I., 'Human, all too human: accounting for automation bias in generative Large Language Models', presented at the ACM Conference on Fairness, Accountability, and Transparency (ACM FAccT 2024), in Rio de Janeiro, Brazil, 3-6 June 2024, published in International Data Privacy Law.
Carnat I. 'L'autonomia del danno morale: una cura con il contagocce' (2024), nota a sentenza Cass. civ. sez. III, 28 novembre 2022, n. 34927, in Danno e Responsabilità 1/2024.
Carnat I., 'Ethics Lost in Translation: Trustworthy AI from Governance to Regulation' (2023) in Opinio Juris in Comparatione 4/2023.
Carnat I., 'Il compromesso di ChatGPT per il contenzioso civile: dall'affidabilità alla responsabilità' (2023) in Rivista Italiana di Medicina Legale 2/2023.
Carnat I, ‘The Notion of Defectiveness Applied to Autonomous Vehicles: The Need for New Liability Bases for Artificial Intelligence’ (2020) in Trento Student Law Review Vol.2 No.2.
Courses
- Panelist at the roundtable session: AI in the courtroom - different perspectives, 9 December 2024, European University Institute, presenting the ECJ SCHUFA case.
- Member of the organizing committee of the 1st Doctoral Workshop on Law, Society, and Artificial Intelligence, 2 December 2014, Pisa, Italy.
- Speaker at the Lawtomation Days 2024, IE Law School, Madrid, Spain, presenting the paper "Addressing the Risks of Generative AI for the Judiciary: the Accountability Framework(s) under the EU AI Act".
- Speaker at "Le regole dell’Intelligenza Artificiale" con "Umana, troppo umana: l'intelligenza artificiale generativa nei processi decisionali" (Università di Pisa, Facoltà di Giurisprudenza, 15 April 2024)
- Speaker at "Intelligenza Artificiale: opportunità o minaccia per l'avvocato" (Camera Civile di Arezzo, 9 February 2024)
- Discussant at "Digital Justice and Constitution" (European University Institute, Florence, 19 January 2024)
- Speaker at "Responsabilità da Intelligenza Artificiale: profili etici e giuridici" (Milano Digital Week, 9 October 2023)
- Speaker at "Come l'IA cambia le professioni legali" (Milano Privacy Week, 28 September 2023)
- Teacher at “Strategia europea dei dati e le ricadute du PP.AA. e aziende” with “La sfida dei dati aperti e il ri-uso dei dati: il caso della giustizia predittiva” (AILUN, Nuoro, 21 June 2023)
- Speaker at IAIL 2023 "Imagining the AI Landscape after the AI Act" presenting the article "Holding the AI Act Accountable" (Monaco, Germania, 27 June 2023)
- Speaker at “ADELE final conference” presenting the Predictive Justice Project (European University Institute, Florence, 9 June 2023)
- Speaker at "Giustizia e algoritmi" (Ordine degli Avvocati di Massa Carrara, 17 March 2023)
- Speaker and co-organiser of “Battle of Brains”: Debate Workshop (Università di Pisa, 6-7 December 2022)
- Speaker at “L'algoritmo incontra la Giustizia: nuove sfide per l'Avvocatura” (Ordine degli Avvocati di Prato, 4 December 2022)
- Speaker at “No al giudice robot, sì alla tecnologia al servizio dei diritti” (Milano Privacy Week, 28 September 2022)
- Speaker at “Intelligenza Artificiale, Diritto e Sci-Fi” (Fondazione Scuola Forense Alto Tirreno, 24 June 2022)