Abstract: 

Artificial intelligence (AI), as most new technologies during human history, is a double-edged sword. It is well-known, that AI can pose risks to society or individuals if it was not designed with care and checks. In this workshop, we will focus on AI risk to companies. Only a few business leaders have had the opportunity to sharpen their intuition and adapt their organization governance framework to track and mitigate these risks, starting from in-house AI-development efforts and finishing with evaluating and monitoring third-party AI tools.

By completing this workshop, you will develop an understanding of various vectors of AI risks to companies and be able to improve the governance inside your organization. Additionally, you will get hands-on experience on the problem of biased data and simulate an adversarial attack on a neural network model.

Session Outline
Lesson 1:
In the beginning, we will familiarize ourselves with AI risks to companies, exploring different types of risks, such as lack of trust, bias in data, automating unethical actions, lack of monitoring and how how to setup the governance framework within your organization.

Lesson 2:
AI uses large volumes of data and then, using algorithms, learns on training data from the patterns and takes actions on previously unseen data. We will explore hands-on, how AI produces faulty results when the data is biased or problematic.

Lesson 3:
In the last lesson, we will simulate adversarial attacks, where we as the attacker will find a set of subtle changes to an input that will cause the target model to misclassify and provide the output we want.

Background Knowledge
Python programming language for lessons 2 and 3
Knowledge of neural networks with Keras and TensorFlow for lesson 3
Python environment installed (e.g. Jupyter Notebook, Databricks, PyCharm,...)

Bio: 

Lukas Csoka, working as Head Big Data Foundations in Swiss Re, one of the biggest global reinsurers, combines data science, consultative thinking and business partnering as the main skills used daily in his job to push design and develop new AI products. His MS in software engineering from Slovak University of Technology and MBA in global management from City University of Seattle are supporting him during this journey. His projects include innovative solutions for farmers using satellites, the application of predictive methods to manage costs inside Swiss Re or the analysis of thousands of client meetings from text and the provision of recommendations to Swiss Re's traders and many others.