Portal de Periódicos da CAPES

Role of AI and Its Impact on the Development of Cyber Security Applications

2023; Springer Nature; Linguagem: Inglês

10.1007/978-981-99-2115-7_2

2191-6861

Aarthi Iyer, K. S. Umadevi,

Network Security and Intrusion Detection

With the Industry 4.0 revolution, modern organizations and their ecosystems are evolving rapidly. The cyberattack surface is growing simultaneously, thus enabling new cyberattacks that cannot be resolved by human intervention alone. Artificial intelligence (AI) has now become an essential tool that provides security as it has the potential to easily analyze billions of data and triangulate an array of possible cyberattacks that can range from active attacks like network exploitation or suspicious behavior such as eavesdropping, traffic analysis. An AI continues to learn from datasets, improve its models based on historic data and use those experiences to better recognize different types of attacks. Artificial intelligence has a great impact on the cyber security industry by predicting and preventing attacks before they even occur. In the Tokyo Olympics 2021, 450 million cyberattacks were attempted during the entirety of the events, but with the help of AI, none of the attacks affected the tournament because AI predicted attacks a week before it happened. AI has been helping with improved cyber security which would not have been achieved by human component alone. A recent survey done in IBM and Capgemini shows that 95% of security breaches arise due to human error alone. Forty-eight % executives in Capgemini use AI for threat detection, 18% say that it is used for response and 34% make use of AI for prediction. In the era of ever-evolving attacks and rapid increase in number of devices, AI and machine learning can aid in keeping track of attacks, help in threat detection automation, provide effective response to data exploitation, conduct network vulnerability surveillance, generate effective incident and response reports, form cyber threat intelligence reports, which are a few applications of AI in cyber security. As with every technology, AI comes with certain limitations too. Though it contributes to a high degree of accuracy and machine performance, there still exist certain levels of risk like producing false positives and negatives while detecting threats in a network. Using adversarial machine learning techniques to exploit the dataset of an AI and train it to mislead the application into believing threat activities are normal data is a possibility.