
AI/Security
Sanjay Goel
Director of Research at the New York State Center for Information Forensics and Assurance
His research interests include information security, cyber warfare, music piracy, complex systems, security behavior, and cyber physical systems. His research on self-organizing systems includes traffic light coordination, smart grid and social networks. He is actively engaged in policy efforts on cyber security norms, CBMs, and cyber treaties.
Abstract of Keynote Speech
25. 10. 2023
Artificial Intelligence and Society: The Beautiful Promise and the Dystopian Threat
Artificial Intelligence has become the next bastion of technology that will bring a new era of prosperity in society with self-driving cars, smart devices that will manage health conditions, and generative AI that will automate mundane intellectual tasks. As AI becomes more potent, are we ready for the singularity event where a machine becomes more powerful than a human brain. While AI portends a bright future for the society, there are a lot of challenges that need to be addressed as we embrace AI in society. First AI is based on past data and any biases in society encoded in the data can become embedded in AI systems. Second, AI systems can be misused by hackers and criminals for gathering intelligence and facilitating attacks and perpetrating fraud. There are threats to AI systems themselves through various attack vectors including, data poisoning, malformed input, and model tampering. Additionally, there are a lot of societal challenges that come with the integration of AI systems in society. A lot of jobs will be displaced by AI systems, and we need to ensure that we are able to recalibrate our workforce by defining alternate employment for hundreds of million people. AI systems will concentrate power and wealth with companies that control (and incorporate) AI systems which will be less dependent on workers and be able to operate with a much smaller workforce. AI will also be used increasingly in military systems with killer robots and drones as well as sophisticated missiles that can hit targets precisely. Privacy will become a large challenge both through leakage from training data and reconstruction from trained models. Finally, face and voice recognition will make surveillance very efficient allowing governments to be able to track citizens in real-time diluting individual privacy or right to be left alone. In this talk, we discuss the future of society with AI and the threats and challenges that come with it.