The Singularity II: Unguentum Ex Machina
In 2019 we fed the entire history of BPAL scent names into a neural network; from these, it was able to dream up an entirely new batch of products such as “Bayou Fluid,” “Harlequin Milk,” and “Mother’s Hot Ghosts.” Twenty of these perfumes were released as THE SINGULARITY, our first creative collaboration with artificial intelligence.
A lot sure has happened since then!
Now that the proliferation of A.I. has begun affecting workers in every industry and people from all walks of life, we can honestly say we have no idea what the future will look like. But we can offer a preview of what it might smell like, because we’ve done it again! *mad scientist cackle*
Once again, human decisions have been removed from the scent brainstorming phase. Our favorite (human) artist Drew Rausch has dutifully taken these prompts and run wild with them for our label illustrations; our perfumer (also human) has similarly pulled out all the stops, channeling these virtual scents into smelly reality at a geometric rate.
The result is The Singularity II: Unguentum Ex Machina, a fresh batch of oily, piping-hot dystopian weirdness that actually supports living artists… while still keeping us on our wrinkly human toes, creatively speaking. So, everyone wins for now?
Humans also wrote all of the descriptions, and will be bottling, labeling, and shipping the actual product. However, we did ask OpenAI’s ChatGPT to explain why humans will always remain uniquely qualified in the field of fragrance artistry, and here’s what it said:
Relying solely on Artificial Intelligence (AI) to determine what humans should smell like is a bad idea for several reasons:
Lack of understanding of human sensory experience: AI is an excellent tool for analyzing data and recognizing patterns. However, it lacks the ability to understand human sensory experiences fully. Smell is a complex sense that is deeply intertwined with our emotions, memories, and experiences. AI may be able to identify common scents that humans find pleasant, but it cannot capture the nuance and complexity of individual experiences and preferences.
Cultural and regional differences: Smell preferences can vary widely depending on cultural and regional differences. What smells pleasant to one group of people may be unpleasant or even offensive to another. AI algorithms trained on data from one region or culture may not be applicable or accurate for other populations.
Lack of creativity and innovation: AI algorithms are trained on existing data, which means they can only generate ideas based on what they have already learned. They cannot create entirely new scent profiles or take into account new trends or emerging preferences that have not yet been captured in the data.
Ethical concerns: Using AI to determine what humans should smell like raises ethical concerns about the commercialization and manipulation of human sensory experiences. It may also perpetuate unrealistic and harmful beauty standards that prioritize certain scent profiles over others.
In summary, while AI can be a valuable tool for many applications, relying on it exclusively to determine what humans should smell like is not advisable due to its limitations in understanding human sensory experiences, cultural and regional differences, lack of creativity and innovation, and ethical concerns.
Disturbingly well said! Thanks, HAL!