Following my debut AI film “Cliff’s Edge: Shattered,” I wanted to dive back into similar themes of isolation and confrontation with “ECHO.” This short piece tells a cyclical story: a figure desperately chases someone along a dramatic cliff edge, only to reach the end and realize the person they were pursuing was themselves, trapping them in a loop. For the visuals, I intentionally worked with the kind of imagery the AI model often defaults to when generating concepts like moody environments or solitary figures. Leveraging these common visual outputs helped me quickly establish the short’s specific atmosphere and style.

Creating “ECHO” was also an exercise in resourcefulness and efficiency. Built using text-to-image (txt2img) generation followed by image-to-video (img2vid) techniques, it was a self-funded project operating on a tight budget. This constraint pushed me to refine my workflow, aiming for a practical success rate of around 30-50% for usable AI generations. It was a challenge that further honed my skills in prompt engineering and navigating the practicalities of AI filmmaking, reinforcing my excitement about using these tools to bring complex visual ideas to life even with limited resources.