This company, led by a former Google researcher, aims to provide computers with a sense of smell through this method.

One company aims to employ technology to digitize the sense of smell, which AI already mimics sight and hearing.

Osmo, a startup founded by Alex Wiltschko, utilizes artificial intelligence technology to enable computers to generate smells, similar to how they generate images and sounds, as stated on the company's website.

Wiltschko has been "passionate about understanding smell" for a long time, he tells CNBC Make It. "It's this very powerful emotional sense, yet we know so little about it."

He earned a Ph.D. in olfactory neuroscience from Harvard University after studying neuroscience at the University of Michigan and earning a bachelor's degree there.

In the following year, he was appointed as a research scientist at Google Research, where he spent five years heading a team that employed machine learning to predict how various molecules smell based on their structure.

Osmo, initially a research project at Google, was later co-launched as a separate startup in 2022 with the support of Lux Capital and Google Ventures.

As the CEO of Osmo, he states that the company's objective is to enhance human well-being and happiness through the digitalization of our sense of smell.

What is Wiltschko's perspective on how humans can benefit from computers processing scent, how Osmo developed its unique technology, and what he envisions the technology achieving in the future?

How AI scent detection and creation can help humans

What is the primary reason Wiltschko gives for allowing computers to smell? He argues that it is crucial for medical professionals to detect diseases.

"Although we know that odors carry disease detection information, computers are not capable of comprehending and interpreting this data at present."

DON'T MISS: How to be more successful with your money

His long-term goal for the company is to make safer, more sustainable aroma molecules for fragrances in everyday products like perfume, shampoo, insect repellant, and laundry detergent.

"Wiltschko believes that fragrances in products are often created by a small number of secretive companies and that they can be improved by using safer and non-toxic ingredients that don't cause skin or eye irritation."

How Osmo uses AI to digitize scent

At Google Research, Wiltschko's team employed machine learning software to create a "principal odor map." They trained their AI model on a dataset of 5,000 aroma molecules across various categories such as floral, fruity, or minty.

Computers struggle to analyze molecules because of their intricate structures, as discovered by Wiltschko.

"Because one tiny change in a molecule's bond can drastically alter its scent from roses to rotten egg," he explains.

The AI model was able to accurately predict the odor of other molecules by using its knowledge of patterns in the different structures of molecules.

He claims that it had superhuman abilities to predict scents.

Building Osmo's AI model from the ground up

When building their AI model, Wiltschko and his team faced a challenge in finding a comprehensive digital library of scent information, as it wasn't readily available on the internet at the time.

""We discovered that we couldn't rely on anyone else's data after spending a year working with fragrance industry companies that believed they had excellent datasets, but we found to be inadequate," he explains."

Wiltschko and his team constructed "a new type of data," he explains.

Master perfumers provided thousands of molecules and their corresponding scent descriptions, which were then fed into graph neural networks (GNNs) for analysis. GNNs are a type of machine learning algorithm that can detect and analyze relationships between data points, similar to how a social network shows people and their connections through friendships.

Wiltchko's team could utilize GNNs to aid their AI model in comprehending atoms, the connections between them, and how the molecular structure influences its scent, according to him.

What's next for Osmo

Wiltschko states that Osmo's ultimate goal is to use its technology to teleport scents by digitizing a scent in one location and recreating an exact copy in another location.

If you can't recreate the smell of the original, then you're fooling yourself, he says.

Wiltschko aims to continue working towards realizing his long-term goal of utilizing technology to detect diseases earlier.

"Although it won't happen this year or anytime soon, we're on our way to building technology that can detect disease with scent."

To become more successful and confident with your money, enroll in CNBC Make It's new online course. Our expert instructors will guide you in mastering your money and discovering practical strategies to boost your savings, reduce debt, and grow your wealth. Start your journey to financial freedom today with a 30% discount using the coupon code EARLYBIRD until Sept. 2, 2024.

Sign up for CNBC Make It's newsletter to receive tips and tricks for success at work, with money and in life.