After Lucknow Police announced that they will equip the city with 'Facial expression recognition' enabled cameras, many citizens have raised concerns about privacy, safety, and efficiency of AI-enabled cameras which will click pictures of women in distress and alert the nearest police station.
Facial expression recognition software is a technology that confirms an individual's identity using their face. It can be used to identify people in realtime. This technology works as a sentiment analysis tool and detects six basic universal expressions -- happiness, sadness, anger, surprise, fear, and disgust.
Working of Facial Expression Recognition Cameras
AI-enabled cameras use algorithms to instantly detect faces, code facial expressions, and based on that recognise emotional states. Facial technology systems tend to operate as follows:
Step1: Face Detection
The AI-enabled camera detects the shape of facial components and captures it. The captured image may show different profiles of a facial structure.
Step 2: Face Analysis and Emotion Classification
The captured images are analysed and the changes in facial features are classified into emotions. These changes are -- the distance between your eyes, the depth of your eye sockets, the distance from forehead to chin, the shape of your cheekbones, and the contour of the lips, ears, and chin.
Step 3: Conversion of Emotions into Data
The captured information is automatically converted into a set of data using a mathematical formula. This numerical code is called faceprint.
Step 4: Ouput Received
Based on the data obtained, the camera will generate an output to indicate whether the image captured of a person is angry, sad, or being attacked, and alert the nearest police station accordingly.
Scope for Error
The data obtained can never be 100 percent accurate, which could lead people to be implicated for crimes they have not committed. "No machine is perfect and the data obtained can definitely display a 'false positive' error," Cyber Security Expert Rajshekhar Rajaharia told The Quint.
Restriction on Privacy
The question of ethics and privacy is the most contentious one. Many citizens worry that the use of AI-enabled cameras could restrict individual freedom."These systems are highly intrusive because they rely on the capture, extraction, storage or sharing of people’s biometric facial data - often in absence of explicit consent or prior notice," read a statement from Privacy International.
Countries that Banned AI Cameras
According to Surfshark, In May 2019, San Francisco became the first city in North America to ban facial recognition technology. Since then, several other cities -- including Oakland and Northampton -- have voted to ban the technology. Following their footsteps, France and Sweden recently banned the use of facial recognition in schools.
Limitations to the Technology
According to Rajaharia, some of the complications to this technology are:
- How will the cameras track expressions while wearing masks is still mandatory?
- What will happen to women wearing a burqa and other facial veils?
- How will the camera differentiate between an argument and an attack?