AI Multimedia: Image Generation & Video Generation

Host Caelan Huntress opened with a compelling parallel to photography's disruptive arrival in the 19th century — when painters felt threatened by a machine that could capture reality at the click of a button. Today, AI image generation is producing the same cultural friction. "Visual art is no longer limited by skill or craft," he argued. "Anyone with an idea can make an image in any style with AI. We are all artists now."

The talk featured a live demonstration of parallel prompting — sending the same prompt simultaneously to ChatGPT, Gemini, and Copilot to compare outputs in real time. The audience collectively chose "a cowboy in Swan Lake" as the test subject, revealing fascinating differences in how each model interprets creative briefs — and where their guardrails kick in (Wolverine and John Wick were off-limits across the board).

Caelan introduced the concept of AI accelerating the five-step creative process — collect, sort, craft, polish, publish — arguing that humans remain essential at the beginning and end: identifying which ideas are worth pursuing, and knowing when something is finished. "It's our creative taste that sets us apart from the machines."

He shared striking market data: in 2024, an estimated 34 million AI images were created every day, and according to PhotoRoom, 71% of images shared on social media are now AI-generated. Jensen Huang's vision of a world where "every single pixel will be generated" rather than rendered felt less like science fiction and more like an imminent reality.

Caelan also introduced OpenClaw — a personal AI assistant for New Zealand businesses — open-source software that runs on dedicated hardware and acts as a 24/7 digital worker. He demonstrated how Puck (his own OpenClaw agent) built openclaws.nz in 40 minutes, and offered OpenClaw installations for Christchurch businesses at a flat rate.

If you want to go deeper on using AI for your business, the AI Coaching Academy offers structured generative AI training in New Zealand for ambitious professionals.

Arthur Machado is an AI filmmaker in New Zealand who works at the intersection of cinematic storytelling and generative AI. At the March meetup he took the baton from Caelan's image generation session and showed how still images created moments earlier could be brought to life — completing the full text-to-image-to-video pipeline in a single evening.

Arthur demonstrated the leading AI video generation tools — including Sora, Runway, and Kling — and walked through the practical differences between them: which excel at smooth motion, which handle camera moves well, and which are best suited to turning a single generated image into a short cinematic clip.

His Image-to-Video cheat sheet (above) distils the key workflow into a practical reference: how to prepare source images for animation, what motion prompts work reliably, and how to chain tools together to extend and refine clips. The cheat sheet was shared with all attendees.

The evening closed with an open Q&A on the ethics of style transfer, the nature of copyright in a generated world, and what it means for creative professionals to have AI tools that can produce polished video from a single sentence.