Can AI help Smoking Cessation Problems?

Can AI help Smoking Cessation Problems?

With news about vaping illnesses and deaths commanding (and rightly so) our attention, the overall ecosystem and interrelationships that helped create this trend are coming under greater scrutiny.

So where did the vaping craze start? For many, vaping is a step-off point for cigarettes, and is one of several tools for smoking cessation, along with gum, patches, inhalers and behavioral therapy.

Vaping has physical actions similar to smoking, (lifting the device to the mouth, inhaling, etc.) which act as a mental surrogate for the act of smoking a cigarette, without the tar or heat. This ‘mental’ element has become a point of import as vapers look to minimize or even quit the habit in light of increased health risk data.

But before we can really talk about the ‘mental’ implications that vaping holds over vapers, we need to understand Cognitive-Behavioral Therapy (CBT). CBT, which aims to challenge and change destructive thought and emotional patterns, has been shown to be effective in helping people quit as part of a comprehensive cessation program. A key element in implementing CBT is identifying and managing triggers so that new, healthy behaviors can be created and smoking related urges can be controlled.

Enter ZiRO, a K8 portfolio company. While I can’t divulge the details on exactly WHAT this exciting new product does, I can hint that it’s going to be a transformative application of artificial intelligence that will help us understand and address behaviors around vaping. Although ZiRO can ultimately help vapers manage and even quit their habit, its technology is really representative of how AI can help us further understand and address complex human issues.

A keystone tenet of ZiRO’s technology and efficacy, is that identifying and breaking key triggers is essential to cessation. To that end, ZiRO incorporates a state of the art deep learning algorithm to automatically understand users’ habit patterns and adapt to changes in those patterns over time. Specifically our implementation of AI uses a Recurrent Neural Network (RNN) with Long Short Term Memory (LSTM) to recognize these habits and triggers, and then distill and present them to ZiRO’s users so they can obtain better insight into the why of their cravings and urges. LSTM allows ZiRO’s RNN to gather data (prior behaviors like how many puffs, when puffed, etc.), think about them in aggregate, and identify correlations between elements, that when taken together, are a trigger for vaping.

Time, for instance, may be a latent trigger leading to a habit. Let’s say the amount of time that passes between puffs throughout the morning leads to a vaper’s greater puff-count density later in the day. A user that takes a series of five puffs every 15 minutes in the morning, may be likely to take 50 puffs every 15 minutes, six hours later. The morning puffs, in this example, act as a trigger for the habit that occurs later in the day. ZiRO’s model would identify this correlation, due to the time series nature of the LSTM, it utilizes. Therefore, the morning puffs would be analyzed as the features that predict the habit later in the day, giving the vaper insight into how their habits are formed on a more macro level.

Armed with this insight, ZiRO users are in a position to manage their own behavior. If you cannot measure it, you cannot manage it. We want to go beyond that adage, by visualizing habits to provide ZiRO users with measurement, synthesis and, most importantly, control.

Technology, if judiciously applied, is a force of good. Will ZiRO be the panacea for smoking and vaping abatement? Definitely not- there are very few silver bullets in life, mostly because what ails us isn’t a result of one single thing, but rather a confluence of multiple factors. So the solution cannot be one thing, but a convergence of multiple capabilities. So, no, ZiRO, won’t be a simple, one stop solution to quit vaping. But that’s not it’s purpose. ZiRO is intended to find the intersection of what makes quitting so difficult — by revealing hidden triggers and habits. ZiRO gives its users clarity and insight that they can use to change their lives, breath by breath.


As mentioned in earlier blogs, K8 is building a ‘skill tree’ where each company we create builds our expertise and ability to unlock new opportunities. With ZiRO we’re leveraging the AI expertise we earned in P3rceive (which is looking for beta customers now- shameless plug), and developing new skills in IoT, manufacturing, packaging and branding. Our partners on this journey are mHub, Stevens and Tate, and Devscale.

ZiRO will be available in 2020.