Text-to-image generation is a design tool that employs artificial intelligence to create imagery from natural language prompting. While the use of text-to-image platforms is gaining momentum in design fields and has the potential to disrupt non-generative design processes, there is limited analysis of the technology in relation to landscape architecture and allied fields. Thus, this paper explores how text-to-image platforms can be used in the design process, what its current limitations are, how its outputs compare to those produced without artificial intelligence, and if people can identify machinegenerated outputs. To do this, the research team tested three text-to-image platforms, created a series of renderings from the three platforms using the same prompt, and found similar images generated without artificial intelligence. The team then set up an exercise to see if people could correctly identify which images had been machine-generated to better understand the current capabilities of the programs. In analyzing the results from the exercise, the team found that participants had a difficult time distinguishing between machine-generated images and human-generated images as only 43% of the guesses were correct. The team also found that while text-to-image tools have the greatest potential when used on the front-end of a project when quick ideation and iteration is key, the tools currently face a number of limitations including reductionism and inappropriate detail capture.