Generative AI as a human-centered visualisation tool | Hao Wei Tu | TEDxSwinburne University

Quick Overview

Hao Wei (Howard) Tu argues that generative AI tools, like those based on Stable Diffusion and Midjourney, should be used not just for creating novel or silly images, but as a bridge to help humans visualize and communicate abstract concepts, feelings, and the long-term future, overcoming the inherent limitations of language and mental barriers to foster deeper collaboration and understanding.

Key Points: The speaker explores using generative AI tools (like GPT, Midjourney, Stable Diffusion) to visualize abstract concepts such as loneliness, quantum physics, or future scenarios. The goal is to overcome the limitations of human minds and language by providing tangible, visual representations for these abstract ideas. A key example shown is an AI-generated image of a 2050 medical scanner shaped like a sphere, designed to give a warm, caring feeling through a soft, heat-retaining cloth, contrasting with current, intimidating scanners. The speaker suggests that AI-generated visuals act as a 'bridge' or 'scaffolding' to connect diverse professionals (scientists, researchers) who may struggle to communicate abstract concepts like 'visions' or 'emotions' effectively. The presentation challenges the audience to move beyond using AI for simple, fun creations and instead leverage it to articulate complex, often ineffable, realities. The core message is that AI visualization can help bridge the gap between abstract internal feelings/visions and external communication, making the invisible visible.

Context: Hao Wei (Howard) Tu, a PhD Candidate at Swinburne University, presents at a TEDx event focusing on the potential of generative Artificial Intelligence (AI) tools beyond mere novelty creation. He positions these tools, specifically mentioning text-to-image generators like those using Stable Diffusion or Midjourney, as essential instruments for improving human communication, particularly when dealing with complex, abstract, or future-oriented concepts that are difficult to articulate solely through words.

Raw markdown version of this recap