Before the seminar, my knowledge of AI was very limited. As a designer, AI seemed a bit beyond my understanding. I perceived AI primarily as a tool capable of creating aesthetically pleasing 3D renders—minimalist and organically beautiful. Over time, AI tools became more familiar to me through friends using chat GPT and the emergence of AI voice cloning in popular songs.
For instance, the song "Saiyan" by l’Heuss L’enfoiré and Gazo (French rappers) was transformed using AI voice cloning of the singer Angèle (Belgian pop star) literally by a guy in his bedroom. The transformed version gathered 1 million views and 102,000 likes on YouTube. This raised interesting questions about the implications and morality of altering artistic content. While it may seem like stealing someone's uniqueness, it also offers a means to preserve specific vocal tones indefinitely. Observing public and artist reactions revealed a surprising acceptance, although concerns about job security and the negative portrayal of AI in the arts persist.
The seminar expanded my understanding of AI's versatility beyond my initial perceptions. I had not previously encountered the 'black box concept,' which metaphorically describes the lack of transparency in an AI system's decision-making processes. Despite its seemingly magical complexity, I learned that these technologies are grounded in scientific principles, intricate algorithms, and advanced engineering. Creating our AI bot for food waste during the seminar felt empowering, providing a sense of control and understanding that made AI more approachable.
The complexity of latent spaces also fascinated me. The term "latent" refers to something that is present but not visible, and a latent space captures underlying patterns or features in the data. The digital visualisations we explored revealed a profound level of entanglement, like the awe-inspiring feeling when contemplating the vastness of the universe. While the topics covered in the MDEF course initially scared me a bit with their complexity, I am learning to embrace this complexity, recognising that certain aspects may challenge the human brain's understanding.
Mobirise.com