How Facial Recognition & AI Are Used to Monitor Smoking in Schools

Student vaping has exploded in recent years, far outpacing traditional cigarette smoking. According to the 2023 Ontario CAMH OSDUHS study, 13.4% of middle and high school students now vape daily [citation:9]. In Windsor-Essex, 5.6% of students in grades 7-12 report daily vaping — compared to just 0.6% who smoke tobacco cigarettes daily [citation:3]. This dramatic shift has prompted school boards to invest heavily in technology to catch students in the act.

The most widespread technology being deployed in Canadian schools is the vape detector — a small device installed on bathroom ceilings that functions like a smoke alarm for vape aerosols [citation:3].

📌 How They Work:

• Particle detection: The sensor detects airborne particles unique to vape aerosols and cigarette smoke [citation:9].
• Real-time alerts: When vaping is detected, the system sends an instant text message or email alert to school administrators [citation:3].
• Location tracking: Alerts include which bathroom and specific time of the incident [citation:9].
• Noise monitoring: Some models (like Piera’s Canaree IX6) also detect loud, sustained noises, alerting staff to fighting or other disturbances [citation:9].

📌 Privacy Protections:

To address privacy concerns, the Ontario Ministry of Education requires that vape detectors do not have audio or video recording capabilities [citation:3]. However, many schools pair these detectors with external CCTV cameras positioned outside bathroom entrances to identify who entered during the alert window [citation:3].

In March 2024, the Ontario government announced a $30 million CAD investment over three years to install vape detectors, security cameras, and other safety equipment in public schools [citation:6][citation:8].

📌 Key Details:

• Funding period: Three years (2024-2027) [citation:8].
• Vendor of Record: Piera Systems Inc. (Mississauga, ON) selected by Supply Ontario to supply Canaree IX6 Vape Detectors with Noise Monitoring [citation:9].
• Technology capabilities: The Canaree IX6 uses AI to distinguish vape and smoke from other particles (e.g., dust, cooking), reducing false alarms [citation:9].
• Integration with security systems: Data from detectors can be integrated with existing video security systems to identify repeat offenders [citation:9].

📌 Education Minister’s Statement:

Beyond simple vape detectors, schools and universities are exploring computer vision systems that can automatically identify smoking behavior through existing CCTV cameras.

📌 How AI Smoking Detection Works:

• Object detection algorithms (YOLOv8, Faster R-CNN, SSD): These AI models are trained on thousands of annotated images to recognize cigarettes, vapes, and smoking hand gestures [citation:2][citation:5][citation:7].
• Pre-smoking detection: Advanced systems can identify when someone is handling a cigarette before they light it, based on hand position and object recognition [citation:10].
• Real-time alerts: When smoking is detected, the system sends instant notifications via WhatsApp, email, or SMS with timestamp and snapshot [citation:7].
• Facial recognition integration: Some systems can be integrated with facial recognition databases to identify students by name [citation:10].

📌 Research Findings:

A 2025 Indonesian university thesis compared two AI detection algorithms for campus smoking enforcement. The study found that Faster R-CNN achieved 100% accuracy in detecting smoking activity at a 0.5 threshold, while SSD only reached 42.5% [citation:5]. The researchers concluded that AI-based computer vision is highly effective for enforcing no-smoking policies in educational settings [citation:2][citation:5].

The Windsor-Essex Catholic District School Board (WECDSB) piloted vape sensors in three schools before expanding district-wide. Here’s what they found [citation:3]:

• Pilot results: While the board has not released hard data, principals reported that the detectors had a “significant effect” on reducing vaping in bathrooms [citation:3].
• Expansion: Sensors are now in 20 washrooms, with plans to reach all 76 high school bathrooms by December 2025 [citation:3].
• Funding received: $89,000 in 2025, with double that amount anticipated over the following two years [citation:3].
• Disciplinary approach: Students identified through detectors and external cameras face progressive discipline — starting with conversations and education, with suspension as a last resort [citation:3].

While school boards argue that vape detectors are necessary for student health, privacy advocates raise serious concerns about the normalization of surveillance in educational settings.

📌 Key Concerns:

• Cameras outside bathrooms: While the detectors themselves don’t record video, pairing them with external CCTV cameras creates a de facto identification system that can track every student entering a bathroom [citation:3].
• False positives: Vape detectors can be triggered by steam, dust, or aerosols from personal care products, potentially leading to wrongful accusations [citation:9].
• Facial recognition concerns: If AI-powered smoking detection systems are integrated with facial recognition databases, students could be tracked across campus without their knowledge or consent [citation:10].
• Normalizing surveillance: Critics argue that constant monitoring teaches students that privacy is not a right — preparing them for a world of corporate and government surveillance [citation:4].

📌 Current Protections:

The Ontario Ministry of Education has stated that vape detectors installed with provincial funding “are required to not have audio or video capabilities” to protect student privacy [citation:3]. However, this restriction does not apply to the external CCTV cameras that schools install separately — nor does it regulate AI-powered systems that may be deployed in common areas like hallways and cafeterias.

• Expansion beyond Ontario: Other provinces are watching Ontario’s pilot programs closely. Quebec and British Columbia have expressed interest in similar initiatives.
• Integration with attendance systems: Future systems could automatically flag students who are frequently detected in bathrooms during class time.
• Predictive analytics: AI could identify students at risk of smoking based on behavior patterns, enabling “preventive” interventions — but raising significant civil liberties concerns.
• Provincial privacy legislation: With no federal law specifically regulating AI in schools, provinces may need to develop their own frameworks for transparency, consent, and appeal rights.