AIAI Seminar - 20 June 2022 - Talks by Claire Barale and Tarini Saka

Speaker:  Claire Barale

Title:

Enabling ethical human-AI reasoning in international law

Abstract: 

AI-powered legal system have been researched as early as the 50s and applications in the legal domain are broad and diverse. However little work has been done in international law. My research focuses on asylum decisions and convention refugees. I am looking at supporting all parties involved in refugee status determination in making better decisions by using machine learning tools.

I am applying information retrieval, legal prediction and similarity analysis on the text of the decisions rendered by Canadian courts in the past 20 years. A second important component of the research is to introduce human-centered computing design methods applied to LegalAI to monitor the impact of the system on its users.

Speaker:     Tarini Saka

Title:

A Collaborative Human-AI Approach to Tackle Phishing Attacks

Abstract:

As people increasingly rely on the Web to carry out their transactions, internet frauds such as phishing attacks have become a significant threat to people’s privacy and security. Most organizations consider phishing a major threat and employ a variety of safety measures to defend their organization, this includes the practice of reporting any suspicious emails to the organization's IT help desk. Unfortunately, IT help desks receive many such reports on a daily basis, and handling each case appropriately requires a lot of time and resources. This causes a strain on the system and leads to users receiving a slow and often generic response. In this talk, I will examine the use of unsupervised clustering algorithms to create an intelligent support tool that can group together phishing emails of similar scams, that can be handled together and responded to with similar advice, and hence help IT staff provide timely phishing advice to the users more efficiently.