AiOfAi: 4th Workshop on Adverse Impacts and Collateral Effects of Artificial Intelligence Technologies 

Accepted Papers

• Paper 1: Age classification in Child Sexual Abuse Materials 

Michal Kozbial and Mateusz Kowalczyk (NASK National Research Institute, Poland)

• Paper 2: Assessing the Reliability of Persona-Conditioned LLMs as Synthetic Survey Respondents 

Erika Elizabeth Taday Morocho, Lorenzo Cima, Tiziano Fagni, Marco Avvenuti and Stefano Cresci (IIT-CNR, University of Pisa, University of Florence, Italy)

• Paper 3: Prompt Scene Investigation: Uncovering the Evidence in LLMs 

Lidaw Fabrice Ledjaki, Yuan-Chen Chang, Mohamed Loutis and Esma Aïmeur (University of Montréal, Canada) 

• Paper 4: Industrialized Deception: The Collateral Effects of LLM-Generated Misinformation on Digital Ecosystems 

Alexander Loth, Martin Kappes and Marc-Oliver Pahl (Frankfurt University of Applied Sciences, IMT Atlantique, UMR IRISA, Germany, France )

• Paper 5: The More You Say, the More You Risk: Ethical Concerns in Large Language Model Reasoning Frameworks 

Jinman Zhao, Linbo Cao, Xueyan Zhang, Ken Shi, Yining Wang and Gerald Penn (University of Toronto, University of Waterloo, Canada)

• Paper 6: Not Just People Anymore: When AI Agents Become Insiders

Christian Kengne (Polytechnique Montréal, Canada)

• Paper 7: DEMO: Ontology-Based Automated Security Analysis for AI Systems and Applications 

Jonathan Roy, Lina Rashdan and Fehmi Jaafar (Université du Québec à Chicoutimi, The George Washington University, Canada, USA)

Tentative Workshop Schedule

Welcome and Opening Remarks | 13:30 to 13:40

Keynote 1: Generation and detection of ‘Infox’ | 13:40 to 14:20  Jean-Gabriel Ganascia

Session 1: Trust, Ethics & Risks in Human-AI Systems 

• Assessing the Reliability of Persona-Conditioned LLMs as Synthetic Survey Respondents | 14:25-14:40 

Erika Elizabeth Taday Morocho, Lorenzo Cima, Tiziano Fagni, Marco Avvenuti and Stefano Cresci (IIT-CNR, University of Pisa, University of Florence, Italy) 

• Prompt Scene Investigation: Uncovering the Evidence in LLMs  | 14:40-14:55  

Lidaw Fabrice Ledjaki, Yuan-Chen Chang, Mohamed Loutis and Esma Aïmeur (University of Montréal, Canada) 

•  Industrialized Deception: The Collateral Effects of LLM-Generated Misinformation on Digital Ecosystems | 14:55-15:10 

Alexander Loth, Martin Kappes and Marc-Oliver Pahl (Frankfurt University of Applied Sciences, IMT Atlantique, UMR IRISA, Germany, France ) 

• The More You Say, the More You Risk: Ethical Concerns in Large Language Model Reasoning Frameworks  | 15:10-15:25 

 Jinman Zhao, Linbo Cao, Xueyan Zhang, Ken Shi, Yining Wang and Gerald Penn (University of Toronto, University of Waterloo, Canada)

Keynote 2: Are privacy attacks against machine learning really worrying? | 15:25 to 16:10  Josep Domingo Ferre

Session 2: Security, Forensics & Investigation Systems 

• Age classification in Child Sexual Abuse Materials  | 16:10-16:25

Michal Kozbial and Mateusz Kowalczyk (NASK National Research Institute, Poland)

• Not Just People Anymore: When AI Agents Become Insiders  | 16:25-16:40 

Christian Kengne (Polytechnique Montréal, Canada)

•  DEMO: Ontology-Based Automated Security Analysis for AI Systems and Applications  | 16:40-16:50

Jonathan Roy, Lina Rashdan and Fehmi Jaafar (Université du Québec à Chicoutimi, The George Washington University, Canada, USA)

Keynote 3: The Network That Thinks: From Infrastructure to Intelligence | 16:50-17:35 Merouane Debbah 

Closing Remarks | 17:35-17:40