With more than four in ten UK businesses reporting a security breach or attack in the last year, what role could Artificial Intelligence (AI) play in the future of enterprise IT security? Ankur Laroia, Leader Solutions Strategy, Alfresco, discusses the potential deployments and benefits. How ready is the enterprise to trust AI with its data security?
According to 2018 research from the Department for Digital, Culture, Media & Sport, 43% of UK businesses have experienced a cyber security breach or attack in the last 12 months. Given the requirements of the General Data Protection Regulation (GDPR) on organisations to report data breaches within 72 hours, it has never more vital that IT decision makers have complete visibility and control of their networks.
Artificial Intelligence could provide an extra level of support in the fight against data breaches. AI could not only help in the identification and alerting of breaches, but even assist in the prediction and post-event analysis of data breaches.
The potential role of AI in enterprise security
Artificial Intelligence seeks to replicate human behaviour. So, within an enterprise security context this could involve the automation of time-intensive processing work, decision making and, potentially, facial and speech recognition.
Data processing is where AI could have a particularly significant impact. Machines can analyse data points at speed and will be able to spot anomalies in behaviour and flag potential threats when traffic or data strays from the anticipated norm.
Over time, as the machines learn from more and more data sets, they will be able to become even more accurate at spotting anomalies. Machine learning can replace a lot of the tedious work, and rather than have a human make a decision around that data, there could be a machine to assist and make suggestions for the human to act upon. Once the machine is getting to 90% accuracy, there is no less need for human involvement, the machine should be able to operate autonomously.
Businesses can therefore speed up the identification of risks and free up their IT staff to focus on other critical actions.
Is the enterprise ready for AI in security?
Alfresco surveyed more than 300 IT decision makers in the UK and US in 2018 to see how they feel about AI in cyber security.
The study finds that more than a fifth (21%) believe that AI is mature and will play a significant role in enterprise security. In fact, nearly one in ten (8%) already allow their AI-based security systems to take significant actions on their systems without requiring any human validation.
For the remainder, it is largely a picture of an industry working out how AI can assist. Just over half (58%) of IT decision makers believe AI will play a minor role in security as they figure it out. Yet, almost a quarter (22%) do not believe that AI will play a role in security as it is too soon.
IT decision makers on both sides of the Atlantic are largely unanimous (99%) in being open to AI taking independent security actions, with nearly half (46%) being “very open” to the idea of incorporating AI into their security but are not engaged yet.
While a quarter (25%) of IT decision makers say they do not personally trust AI to mitigate security attacks, just 1% believe that AI should never be allowed to take actions without human validation.
The general picture then is that the jury’s out on AI’s involvement in the future of enterprise security, so it will be interesting to learn from the experiences of the 8% that have already trusted AI to make security decisions for them.
Where does the industry go from here?
There will be much more experimentation in this space as enterprises look to create efficiencies in their cyber security. There is curiosity – if not hunger – among IT decision makers to see how AI could help them further improve their security.
Data drives AI. Machines will deep learn as new data sets appear. The most successful AI projects that we will see in enterprise security will be those that are constrained and aimed at solving a specific problem. We are already seeing code becoming less important and models becoming more significant in AI projects, and this could apply equally for security.
Artificial Intelligence is already playing a key role in many elements of the enterprise, especially in document management, so it is only a matter of time before AI becomes a mainstream component of enterprise security. At the outset, human involvement will be critical to the development of AI in enterprise security but ultimately, the enterprise will benefit as staff are freed up from automatable tasks to focus on other key areas to ensure the security of their data and regulatory compliance.