Ngintip.in: Implementation of Machine Learning Based on Google's Teachable Machine and Waterfall Method to Detect Shoulder Surfing Attack

The use of gadgets in various purpose and places can expose sensitive information, such as private messages and bank account information. Moreover, the use in crowded environments raises potential security and privacy issues in the form of shoulder surfing attack. Shoulder surfing attack is a physical security threats carried out by observers who steal passwords from legitimate users through direct observation or video recording of password entry. Research found that 97% of the participants know about shoulder surfing in daily life and 21% confirms it happens regularly. Moreover, 70% commuters are doing shoulder surfing in United Kingdom with some of them see confidential information. This research proposes the use of machine learning based on Google's Teachable Machine to detect and prevent shoulder surfing attack. The development use waterfall method which includes requirements analysis, design, implementation, testing, and maintenance. The data test produces machine learning model with high accuracy rate, reached about 99.34% in the test case. Finally, the machine learning model is implemented on Chrome Extension with given name Ngintip.in.