Artificial intelligence can take on a bevy of roles: friendly chatbot, vacation planner, image generator. As it turns out, computers also make pretty good perfume consultants.

On Thursday, home fragrance brand LAFCO launched a collection of deodorizing room sprays formulated with the assistance of AI. "This is a two-part technology," LAFCO founder Jon Bresler says of the sprays. The first element has nothing to do with AI — rather, it's a odorless solution that targets the five main categories of malodors: skatole (which smells like feces), dimethyl trisulfide (rotting vegetables), ammonia (decomposing waste), isovaleric acid (feet), and thiogeraniol (sulfur), all of which have unique molecular structures.

LAFCO added this stink-busting solution, which encapsulates and dissolves malodorous molecules, to six of the brand's best-selling fragrances: Chamomile Lavender, Fresh Cut Gardenia, Marine, Champagne, Feu du Bois, and Sea & Dune.

LAFCO

Here's where AI comes in. The formula for each room spray was run through an algorithm developed by Cosmo International Fragrances that, according to Bresler, "identifies things that are good and identifies things that are bad in your formula, and recommends changes to them."

Bresler illustrates: "[The algorithm] will say, 'Okay, I've analyzed your Champagne room spray. I've noticed that Compound A will increase the perception of malodor in skatole. So please remove that from your fragrance compound and instead of putting Compound A in, put Compound Z in.'

"So it's a odor-removing technology, assisted by an AI component, that maximizes your fragrance composition to have the best efficacy at malodor elimination when it's used."

LAFCO

LAFCO's AI-assisted room sprays set the stage for the future of fragrance, one in which computers and perfumers can work hand-in-hand to streamline and optimize the formulation process. "Fragrance compounds are quite large. They can have 40, 50, 60, 100, 200 ingredients in them," Bresler explains. "Having an AI program that's able to analyze all of the ingredients will have an effect on the fragrance community."

To the expert, mood-boosting "functional fragrance" is one category that could use an artificial hand. "Say I'm a perfumer and I want to make a fragrance that promotes happiness... I want to make it smell like gardenia," Bresler imagines. "I would not, as a perfumer, probably have the wherewithal to find only perfume ingredients that promote happiness, and then narrow down that class of ingredients to construct something that smells like gardenia. It seems out of the realm of the human brain." The task of sorting through thousands of perfume ingredients is, however, something an algorithm could do in mere seconds.

While the recent AI boom has fueled anxiety that machine may one day replace man, Bresler believes the role of the perfumer is safe. "I don't know that an AI program could do a final construction of a perfume without smelling it," he theorizes.

"I think the perfumer would have to construct the fragrance, and then run it through an AI program to perhaps boost the end result. Your gardenia smells different than my gardenia, that smells different from somebody else's gardenia," he notes, acknowledging that everyone's perception of smell is informed by genetics, as well as conscious, lived experience — impossible for non-sentient computers. "I don't know that an algorithm could create a gardenia. I think it could only analyze it."