Blackjack: A game model for applying AI to cybersecurity

Cyber-attacks continue to threaten organizations large and small. The impacts of a data breach or ransomware attack may have significant and material impacts on both customers and shareholders. To help combat cyber threats, some organizations have started exploring how big data and artificial intelligence (AI) may help to reduce cybersecurity risk. Machine learning algorithms are now common in cybersecurity. We find machine learning offered in more commercial products, from those that are fully integrated into products and require no knowledge of machine learning to those that require rolling up your sleeves to put together the algorithms and perform statistical analysis. Machine learning for cybersecurity has most frequently been applied to detecting patterns that represent attacks. This includes algorithms that evaluate audit log data, that spot anomalies for network intrusion detection systems, and that identify and block malware on computer systems. In some applications, machine learning is used to train models of normal activity on networks in hope of later detecting anomalous events that may represent a cyber-attack.