EDITOR’S COMMENT
Ch
ib,
V
s
ho
ice
op
Pre
a, S
c
i
r
side
nt Middle East and Af
P
eople have written a lot of breathy pieces about ‘Artificial
Intelligence’ (AI) replacing humans in the workplace. While
there is a lot of clever tech involved in, for example, replacing
a cab driver with an automated driving system, at the end of the day
this substitutional use of technology is really just about saving cost
and improving productivity, and is a continuation of a trend that goes
back to the beginning of the industrial revolution. Today’s self-driving
car is tomorrow’ spinning jenny. In that sense it is kind of boring.
Where AI gets really cool in my view is in enabling us to achieve
things that humans can’t do, or at least can’t do well or in any
reasonable amount of time. This often involves spotting patterns in
vast amounts of data and data that is not represented nicely in the
form of a human sense (sight, sound etc). Humans usually are not
that effective at this level.
In the field of cybersecurity for example there has been a huge
shift in the last couple of years to using AI, specifically ‘Machine
Learning’, for predicting if a piece of software is good or bad before
it runs. We have built models that can predict this with incredible
accuracy and with lightning speed.
With cybercriminals using clever automated systems to churn out
300,000 new pieces of malicious software (‘viruses’) every day, there
is no way that human analysts can stay on top of the problem using
traditional anti-virus techniques.
The most advanced models are built using an approach called
‘deep learning’, using ‘artificial neural networks’ based on the
neurons in the human brain but tuned heavily towards this specific
problem domain.
These models are already deployed on millions of computers
in businesses around the world. Deep learning has been incredibly
effective at stopping new malware campaigns, ransomware etc.
We are already working to apply AI/deep learning techniques to
other problem areas – for example predicting if a new web page is
okay or malicious before anyone opens it or predicting if a pattern
of behaviour of a user in a company network is normal or if their
account has been hacked.
INTELLIGENT TECH CHANNELS
Issue 18
ARTIFICIAL
INTELLIGENCE –
A REAL HERO IN
CYBERSECURITY
HARISH CHIB, VICE PRESIDENT MIDDLE EAST AND AFRICA, SOPHOS
In cybersecurity, AI will have a massive
impact on our ability to stop cybercriminal
activity. By using more AI we can give
the overstretched teams working in
cybersecurity in organisations a better
chance of stopping the cybercriminals.
AI is even more helpful in cybersecurity
as there is a massive shortage of trained
cybersecurity professionals globally and that
is not going to change.
Where AI gets really
cool in my view is in
enabling us to achieve
things that humans
can’t do, or at least
can’t do well or in any
reasonable amount
of time.
17