AI in cyber security: friend or foe?
02:00 Monday, 05 July 2021
UK Cyber Security Council
According to a World Economic Forum report published in late 2020, by 2025 “85 million jobs may be displaced by a shift in the division of labour between humans and machines” – that is, automation, Artificial Intelligence (AI) and Machine Learning (ML) will take over the functions of those people. This is a staggering number, but in fact it’s not as bad as it sounds.
First, machines are very good at doing boring, simple work. If you can automate a tedious manual task, it is likely to be done more efficiently and more accurately by a machine than by a person. Take a contact centre in the holiday industry, for example: when they introduced technology that could automatically figure out which advertisement the caller had been reading, they discovered that the manual: “how did you find out about us?” entry was done incorrectly 28% of the time on a good day, 43% on a bad day. Now, in some cases automation will result in a human being leaving the business – but there are two more likely outcomes. First, if you can automate the boring parts of a job that don’t need a human brain, the person who did that task can now do something more valuable. Second, one tends to find that if a person really is no longer needed, the approach tends to be not to replace someone who leaves rather than to show them the door directly.
Second, machines are really good at doing statistical analysis and chomping up massive data sets (which is, in reality, what AI and ML are doing). So, these tools are beginning to do work that we simply couldn’t do before due to the technology not having been invented yet. And with tools such as SOAR (Security Orchestration, Automation and Response) we’re beginning to see the machines take the strain and let the people do the proactive human-preferable elements of threat hunting.
Third, in case nobody had noticed we have a skills gap – more of a chasm, really – in cyber security. There are more jobs than there are suitable people, and even if AI and ML displace some of them, a huge gap will still remain.
And finally, there is a flip-side to the WEF’s estimate of those 85 million jobs ceasing to exist: “97 million new roles may emerge that are more adapted to the new division of labour between humans, machines and algorithms”, says the report.
After all: AI and ML may do some of the boring and/or difficult work in cyber security, but we will need people to produce the AI and ML systems in the first place, and to keep them fed and watered.