Open Source Summit Europe 2022

In this session, open source advocates, researchers and developers will be invited to participate in a thought-provoking discussion about a seismic shift facing open source today: What characteristics should be in an Open Source AI? The traditional view of open source code implementing AI algorithms may not be sufficient to guarantee inspectability, modifiability and replicability of the AI systems. AI is already deciding who stays in jail or which customers deserve credit to buy a house. More kinds of “autonomous” systems are appearing so fast that government regulators are rushing to define policies. The Open Source Initiative is leading an exploration of the world of AI and Open Source, diving deep around the boundaries of data and software to discover how existing concepts like copy, distribution, modification of source code apply in the context of Artificial Intelligence. We’ll share the first results of OSI’s Deep Dive: AI, the multi-part event that is helping OSI stakeholders frame a conversation to discover what’s acceptable for AI systems to be “Open Source.”

Artificial Intelligence (AI) is no longer in the future. AI has left the research laboratories and has become rapidly part of various aspects of our lives. As AI adoption grows and the system's complexity increases, the consequences of making wrong decisions also rise significantly. The more we depend on AI to make decisions on our behalf, the more we require mature AI systems that we can trust. To ensure AI systems continuously behave in a trustworthy manner, research communities and governmental bodies began efforts to define guidelines and principles like Trusted AI, Responsible AI, Ethical AI, and Trustworthy AI. Each guide gives its taxonomy and categories of properties the AI needs to achieve for the good of humanity. In this session, Teodora and Diana will walk you through the Trusted AI ecosystem and demo existing open source projects: Adversarial Robustness Toolbox, AI Fairness, AI Explainability, and more to show how to leverage the existing tools to insert accountability into your machine lifecycle and improve the trustworthiness of your system.

Artificial intelligence is growing into a powerful technology. With enormous power of AI, comes great responsibility. AI raises worries on a number of fronts due to its potential for disruption. There are many concerns like workforce displacement, privacy, potential biases in decision making. Explainable artificial intelligence (XAI) is a set of processes and strategies that allow humans to understand and trust ML algorithms'. XAI is one of the important prerequisites for adopting responsible AI, a methodology for deploying AI approaches in real businesses on a big scale while maintaining justice, explainability, and responsibility. As AI becomes widely used, "adversarial AI" emerges, as attackers take advantage of ML algorithms. Many businesses focus their security expenditures on hardware and software attack surfaces, leaving a critical vulnerability unprotected. "Bias in AI" has been a topic of research and concern in ML area, and it has gained in popularity among broad consumers in recent years as AI understanding has increased. Dr. Vamsi Mohan Vandrangi will be talking on the Trusted and Responsible AI. He will discuss on the Explainability (XAI), Adversarial AI/ML and possible attacking methods, Bias in AI and their fixing methods, and Fairness in AI systems.