This session explores the cutting-edge developments in AI-driven automation of scientific research workflows. Students will examine state-of-the-art systems that can autonomously conduct literature reviews, generate research hypotheses, design and execute experiments, analyze results, and draft scientific papers. We will consider appropriate use cases for automated scientific workflows, identifying research problems that are well-suited for AI automation versus those requiring human oversight and intervention. We will also identify risks in autonomous scientific workflows, including recognizing hallucinations, citation errors, methodological flaws, and other failure modes that compromise research integrity. The session prepares students to navigate the evolving landscape of AI-assisted research across disciplines.