New AI Device Detects Both Vaping and Bullying in Schools

Deep Learning algorithm able to sense e-cigarettes and bullying.

Posted Jun 06, 2019

Source: Anemone123/Pixabay

There are two growing problems negatively impacting America's youth: smoking e-cigarettes (vaping) and bullying. One pioneering startup is applying state-of-the-art artificial intelligence (AI) to help fight both problems. In late March, Soter Technologies based in Hauppauge, New York, introduced their latest cloud-based machine-learning solution that detects and send alerts when vaping or bullying occurs.

Tobacco use by U.S. youth has significantly increased due to vaping—over 3.5 million American middle and high school students used e-cigarettes last year, according to the Food and Drug Administration (FDA). E-cigarettes (electronic cigarettes) are devices that vaporize a chemical solution made of nicotine, flavorings, and other additives in order for it to be inhaled into the lungs. These devices may also be used for other drugs. Vape pens, e-hookahs, vapes, tank systems, electronic nicotine delivery systems (ENDS), mods, e-cigs, are all considered e-cigarettes and may come in various shapes to resemble cigarettes, pipes, pens, cigars, and even USB flash drives.

A recent Stanford University School of Medicine study published in the June 2019 edition of the Journal of the American College of Cardiology, showed that using flavored e-cigarettes may increase the risk of heart disease—specifically, acute exposure to “flavored e-liquids or e-cigarette use exacerbates endothelial dysfunction, which often precedes cardiovascular diseases.”

The Centers for Disease Control and Prevention (CDC) warns that tobacco product use among youth, including e-cigarettes, is not safe for many reasons. For example, there are around 70 chemicals known to cause cancer are in cigarette smoke, and nearly all tobacco products contain nicotine according to the CDC. E-cigarette aerosol typically contains nicotine, and may also contain lead, nickel, tin, volatile organic compounds, diacetyl (a flavoring associated with a serious lung disease), and ultrafine particles that can be inhaled. According to recent CDC figures:

  • The use by youth of any tobacco product jumped by 38.3 percent during 2017-2018
  • 1.5 million more current youth e-cigarette users in 2018 than the year prior
  • Over one in four high school students had used a tobacco product in the past 30 days in 2018
  • 4.9 million youth were current tobacco product users in 2018

According to the CDC, nicotine has been discovered in some e-cigarettes that claimed to have zero percent nicotine. Nicotine use in youth is especially problematic as it can negatively impact brain development, attention, learning, and memory. Nicotine is highly addictive.

The use of e-cigarettes was declared an epidemic last year by Dr. Jerome M. Adams, the 20th Surgeon General of the United States and Vice Admiral in the U.S. Public Health Service Commissioned Corps.

Bullying is common, frequent and widespread in the United States. According to a 2018 CDC report:

  • Nearly 12 percent of public schools report that bullying occurs at least once a week
  • One in five high school students reported being bullied on school property in 2017
  • Bullying is among the most common discipline problems in public schools
  • Over 15 percent of high school students reported being cyberbullied in 2017

Soter Technologies was founded as Digital Fly in 2015 to focus on social media awareness for safer schools and communities to overcome bullying. FlySense is a vaping and elevated sound detector launched in August 2017 to address the teen vaping epidemic in U.S. schools. The latest version, FlySense 255, was launched recently on March 2019. In the last two years, there has been significantly rising demand.

FlySense is an AI cloud-based platform with Wi-Fi sensors that can detect and send alerts with vaping or sounds associated with bullying are detected. Because it is sensor based, it can go into bathrooms, locker rooms, and other areas where cameras for monitoring are not allowed.  It is a deep neural net that was trained and calibrated with proprietary data. The latest version, called the FlySense 255, supports 5 GHz Wi-Fi and is able to detect dangerous Volatile Organic Compounds (VOCs) that are present when vape pens and e-cigarettes are used. The machine learning is an evolutionary algorithm that learns in real-time updates sent over the cloud hosted on AWS. Monitors can report false positive alerts that will instantaneously be input into the AI machine learning solution either over a web portal, using a smartphone app, or by email.

“When we first started, inquiries for our devices were about a couple per week, said Derek Peterson, CEO of Soter Technologies. “Now we get around 60 per week.” According to Peterson, they are producing around 600 devices per month. The patent-pending technology is currently deployed in hundreds of schools in 35 states and 10 countries.

Copyright © 2019 Cami Rosso All rights reserved.


Stanford Medicine (2019, May 27). E-cigarette use, flavorings may increase heart disease risk, study finds [Press Release]. Retrieved from https://med.stanford.edu/news/all-news/2019/05/e-cigarette-use-and-flavorings-may-increase-heart-disease-risk.html

Lee, Won Hee, et. al. “Modeling Cardiovascular Risks of E-Cigarettes With Human-Induced Pluripotent Stem Cell–Derived Endothelial Cells.” Journal of the American College of Cardiology. June 2019.

CDC. “Vital Signs.” Feb 2019. Retrieved June 6, 2019 from https://www.cdc.gov/vitalsigns/youth-tobacco-use/index.html

U.S. Department of Health & Human Services. “VADM Jerome M. Adams, M.D., M.P.H.” Retrieved 6-6-2019 from https://www.hhs.gov/about/leadership/jerome-adams/index.html

National Cancer Institute. “NCI Dictionary of Cancer Terms.” Retrieved June 6, 2019 from https://www.cancer.gov/publications/dictionaries/cancer-terms/def/electronic-cigarette