AI in a Minefield - Learning from Poisoned Data

Explore the challenges and implications of AI systems learning from poisoned data in this 49-minute conference talk from BSidesLV 2021. Delve into the concept of ground truth and its importance in machine learning as Mr. Itsik Mantin discusses the potential risks and vulnerabilities associated with training AI models on compromised datasets. Gain insights into the strategies and techniques used to mitigate the effects of data poisoning attacks and ensure the reliability and integrity of AI systems. Understand the critical role of data quality and validation in developing robust and trustworthy artificial intelligence applications across various domains.