The increase in technology usage among children has brought on new and more complex challenges for many companies in managing the use of their sites and applications to avoid misuse for grooming and sextortion. These companies are constantly searching for innovative solutions for improving their cyber security. Breaches mean a loss of valuable data, money and customer trust. In 2017 the cybersecurity market was worth more than US$105 billion, but it is projected to reach US$259 billion by 2025. Originating from the Norwegian Biometric Laboratory (NBL) at NTNU, a cross discipline team of researchers have developed novel and accurate algorithms proven theoretically, to detect child predators and people with fake profiles online. The algorithms are based on machine learning approaches within the fields of keystroke dynamics and stylometric behaviour and they form the fundament of the AiBA risk analysis solution. The aim of AiBA is to provide warnings of suspicious or potentially harmful activity in real-time. Initial testing has achieved a higher accuracy than state-of-the-art solutions. The commercialisation strategy of the project is to start a venture funded spin-off company that will get access to the technology through an exclusive license agreement. The business idea of AiBA is to offer an accurate and real-time solution for detection of fake profiles and child predators online to social media and gaming companies, through API license agreements. API licenses are a common and well proved model within the industry. We apply for 5 MNOK for verification of the AiBA technology in a real environment with a 3rd party industry player. The multinational Danish gaming company MovieStarPlanet, targeting players age 8 to 13 and with 400 million registered profiles, is interested in AiBA and has signed a letter of interest and wants to contribute with chat data and a pilot test in their system, given that implementation seems viable through funding.
Project leader: Hege Tokerud
Institution: NTNU TECHNOLOGY TRANSFER AS