This session will help users learn tools to better interact with AI systems, informed by how real users have interacted with existing systems, such as ChatGPT. We’ll look at the existing shortcomings of AI systems, including limits of social intelligence, biases, possible pitfalls of relying too much on AI systems' outputs, and negative impacts of this technology. We’ll also discuss how users can brainstorm and avoid negative risks of their own inputs to the system, drawing from our work in understanding negative impacts and risks of AI.