AiThority Interview with Guy Goldman, Founder and CEO at Olive Diagnostics

Hi Guy, please tell us about your journey into AI technologies and how you arrived at Olive Diagnostics?

To start with, the reason the technology was initially developed was due to the fact that while Guy Goldman was managing a company in London, he was informed by his mother living in Jerusalem, Israel, that she was diagnosed with Ovarian Cancer. While she was undergoing chemotherapy Guy was continuously calling home to find out how she was feeling (all he knew about chemotherapy was bad side effects). He was looking for basic information, such as sleeping, eating, drinking, walking etc. His family in Israel made it clear that although it was heartwarming seeing his caring, calling 4 times a day was driving them nuts and it had to stop, either return home, or stop calling but the situation needed to change; Immediately! 

While Guy didn’t want to return to Israel, he started seeking technologies to monitor his mother from a distance and was surprised that nothing existed. At the time the only viable product was a smartwatch that would give just heart rate and maybe steps. Because the lack of technologies in the market, and the need for such technologies, Guy started looking for a solution with his mother as the persona in mind. This meant that the technology needed to be 100% passive, easy to install and easy to use, and most importantly – whatever the technology was, it couldn’t change anything in the way his mom lived her life. 

AI is the required technology needed in order to develop models that can perform molecular detection from the sea of noise that is inherently in the system. In order to understand the need for AI, it is important to understand the noise in our system. We have developed a spectrometer that sits on the side of a toilet. Spectroscopy is a very sensitive technology that is usually performed in a very controlled environment, typically users of spectrometers place a test tube into a control box, that is placed into the spectrometer which is then shut to furthermore shield the sample from external light. Our device is mounted to a home toilet with no protection from the external light. 

Along with that fact is the fact that we are sampling the fluid while it is in movement, the sampling is done while the urine stream is in mid-air, which means it’s moving and doesn’t have any geometrical shape. 

So, the system is filled with noise from the surrounding light, since the urine is moving and that the shape greatly distorts the light passing through the drops. Our system has to cope with all this noise while simultaneously receiving a very small signal. We are flooding a 500 cm² surface area while the signal is at best only 2mm2 so the amount of signal is very very small. All of these conditions forced us to look outside of classical Physics to develop models that can amplify the small signal and reduce to the best of our efforts the tremendous amount of noise. 

AI was a perfect tool for this problem as it allowed us to look at the problem purely from a data perspective as the Physics is too detailed and complicated.

By using AI modeling, we were able to use the data itself with assistance from our physicist to determine the nonlinear characteristics of light passing through a drop. 

Recommended: AiThority Interview with Robert Schena, CEO and Co-founder at Rajant

What is Olive Diagnostics and how it fits into the modern AI ecosystem?

At its most basic sense, Olive Diagnostics has created a very unique sensor that uses some 64 different frequencies and photodiodes to measure the absorbance of light in molecules. The setup requires models that can amplify the tremendous SNR (signal to noise ratio) in the system. The company utilizes AI in many ways, for the most part it is determining the nonlinear relations between the passage of light through drops of fluid. These relationships aren’t defined in classic research literature It is too complicated to develop a unified equation to describe their behavior. AI gives us the ability to extract these equations without having to rely solely on Physics or Chemistry. Models are created from the data itself. 

Furthermore, once the SNR problem is solved, the company utilizes AI to establish models needed to analyze the signal, it is important to understand that the absorption of energy in molecules is so small that it isn’t apparent at all. AI is used to look at thousands of dimensions simultaneously to find behavior that can be traced back to presence of a particular molecule in a sample.       

How do you position Olive WatchOS app in the growing universe of AI-based medical and healthcare personalization apps?

Olive is the only technology in the world providing urinalysis on a continuous basis. No other technology gives this sort of insight. 

By analyzing urine, it is possible to early detect thousands of diseases before symptoms appear. While other Healthcare applications exist, none of them gives such rich data that is critical for monitoring health continuously. Urine acts as a regulator to the body’s internal workings. By analyzing the regulation, it is possible to monitor anything within the body. Most of this stuff hasn’t ever been tested as no one has ever had access to this sort of data. So, while we know of stuff that is commonly associated to urine such as kidney functionality, UTI, Kidney Stones, dehydration etc. There is so much more than can be identified early on through urine. We haven’t even started touching the top of the iceberg. We are looking forward to the studies that are going to be develop by using this technology. 

It is very important to note that it isn’t just medical information that is found in the urine. Nutrition, exercise, general health, safety, law enforcement, fem tech and so many other verticals open up when you have access to urine.  

Recommended: AiThority Interview Series with Eric Visser, President EMEA at GumGum

What are the benefits of using this app? How does it help normal and convalescing patients?

Having simple and continuous access to urine data is key for chronic disease management, as well as verification, to ensure the success of curing. The app makes it even easier for a person to identify themselves to the toilet as well as getting results from the analysis directly to their watch. As a person urinates some 5-7 times a day, he or she can monitor changes that are happening from immediate bodily functions such as the intensity of workout or slower progression situations such as kidney stones or infections.   

Data privacy and data governance are affecting lead times of AI product management lifecycles. How do you foresee these challenges influencing AI medical apps in the US and Western European countries?

The company is focused on privacy and is both HIPPA and GDPR compliant from the ground up. The company is aware and implementing all the necessary requirements to manage privacy in the most secure manner possible. Health data is very private and both HIPPA and GDPR give very good guidelines to what is accepted as a baseline. 

Tell us more about the role of Ethics and AI governance in future applications of data science and IoT.

While the benefits of AI in healthcare are clear, it is important to ensure that there is a separation between the diagnostics of an issue and its curing. By no means should the same IoT device detect the problem and solve it. Olive only performs the data collection and leaves the diagnostics up to other parties. By enforcing this separation, it ensures that a third party validates the findings before administering a cure.  

Your take on the future of AI based diagnostics and how it would simplify healthcare infrastructures.

AI will be a leading technology for future diagnostics! AI simplifies the research in understanding the cause and effect in our bodies, while it can take years to analyze the relationship between different systems in our body it will only take minutes to create prediction models using AI. Olive is doing this in urine and soon in stool but there are many other biomarkers that can be researched in order to predict a person’s health. 

One of the most exciting fields of AI development in healthcare are body degradation models that can predict long term events – when a person is going to lose his or her memory, ability to walk etc. It is clear to all of us that we aren’t going to be playing basketball past our 100th birthday but understanding what our bodies will look like in 20 years can give us great insight into plans we should be implementing now. Live today as if it is your last day is very hard, but having a reasonable understanding about where we are going to be in the mid-term is both a little scary but also a big win to those who are open and eager to know.  

Recommended: AiThority Interview with Amy White, Director of Social Impact and Communications at Adobe

A message to all Healthcare CEOs and Marketers:

It is harder than you initially thought. With that being said, it is surmountable, it is needed and will help people. There is no higher calling to anyone who didn’t learn medicine. Be strong, it too will pass!  

Thank you, Guy! That was fun and we hope to see you back on AiThority.com soon.