Cognitive Computing has just started permeating into our lives and is taking small steps to transform business models in the industry. Its use cases are not limited to speech, image and voice recognition but it also encompasses scenarios in modern medicines, banking and bioinformatics etc. Similarly cyber security is also trying to tap the potential of cognitive computing leveraging Artificial Intelligence (AI), Machine Learning (ML) and Deep Learning (DL). If we rewind our clocks two years back, one can easily identify cyber security as a hot use case discussed in the AI community every now and then. This has raised uncertainties among organizations, i.e. do they have to handle a new born dinosaur or is it something which always existed.
Numerous myths are floating in the cyber security community on the usage of AI, some of them being:
- AI for cyber security is a novelty
- AI for cyber security is one time product fitment
- Cyber security with AI can learn directly from live cases and its data
- Cyber security solutions can only be developed using AI techniques in the future
The realities are, but not limited to, when it comes to cyber security leveraging cognitive computing is not a new concept; rules in security engines have been built for the last 10 years leveraging machine learning techniques. This is because with rise of malware army over the years, it was realized that simplified automation is not a robust solution to thwart advance malwares.
Cyber security experts need to understand that most of the use cases of AI are with static objectives as of today such as facial and speech recognition, it is to be noted that faces and voices are not expected to change. But we all are aware that cyber threat evolution never comes to a standstill. Cyber criminals are always exploring ways to beat the trained defensive models, hence it becomes imperative to continuously update AI rules as configured in the cyber security capabilities. Few more realities which organizations should not ignore are such as, usage of malicious test data to build robust cognitive security stacks and AI based cyber security defence is not a silver bullet, i.e. other solutions are to be leveraged to build robust multilevel protection architecture.
The expectations are high from AI on its usage for cyber security. The key is availability of data from threat intelligence, cybercrime strategies and threat databases etc. Cognitive security is believed to have the potential to optimize security budgets in the long run, enhance incident response, improve accuracy of alerts, and help in staying updated with contemporary threat research and analysis. Let’s us wait, watch, and learn the myths and realities of AI for cyber security as further developments take us closer towards disbursing the real possibilities.