AI can be energy intensive, and artists currently lack access to empowering information. With growing concerns of cli- mate change and calls for environmental sustainability, there is a real need to explore strategies to communicate sustain- ability information to artists using generative AI, given its increasing presence and widening accessibility. This paper presents an exploratory Research-through-Design study (in- cluding design-informing survey, design prototyping, user testing) of integrating sustainability reflection features into generative AI systems, and provides preliminary knowledge of the design characteristics that can be leveraged, including artists’ experiences of them. This paper finds that granular, relatable data visualizations and informed use of colors are effective in communicating about energy consumption. Fur- thermore, artists were positive towards ”feeling bad” in the process of becoming aware of their impacts, and called for systems that could provide them low-energy settings during exploratory stages of the artistic process.