What is Artificial Intelligence?
Artificial Intelligence has been a buzz word in our world for some
time now. Every day thousands of scientists are working on enhancing what
machines can do. Traditionally, machines could not bring about any changes to
the physical world. They could only provide an output, based on the lines of
code a programmer wrote. But those days are long gone.
Imagine a baby being born. How does it adapt to the world around it?
How does it learn? The baby observes and interprets the happenings of the world
around it. Then the baby picks on behaviors which help him accomplish his
goals. The same goes for Artificial Intelligence.
Artificial Intelligence enables machines to gather data from its
activities and analyze them to perform actions. These actions help the machine
reach a step closer to its defined objective. Machines have no sense of what is
right and what is wrong. A machine will only perform its objective and accomplish
what it is meant to do. It all comes down to the human who designed the machine
or the one who is using it whether the person is doing the ethical thing or not.
Though, history tells us a different tale. Those who cannot learn
from history are doomed to repeat it. Signatures have been faked, locks have
been picked, passwords have been brute-forced, networks have been penetrated,
computers have been hacked, biometric systems have been spoofed, credit cards
have been cloned, crypto-currencies have been double spent, airplanes have been
hijacked, CAPTCHAs have been cracked, and cryptographic protocols have been
broken. Millennia!!! long history of humanity contains millions of examples of
attempts to develop technological and logistical solutions to increase safety
and security, yet not a single example exists, where it has not eventually
failed. Artificial Intelligence would make the machines smarter which would
make the world a more secure place but when a human uses that machine for
unethical activities that same security could be compromised.
The Challenges
The total scope of the challenges that can come up in cyber security
is difficult to comprehend at this nascent stage. Yet the signs are there for
all to see. 91% of cyber security professionals are concerned about hackers
using AI against companies in cyber attacks. (Webroot, 2017)
Phishing
Imagine you receive a mail from a friend’s email id. The email is
drafted in a way that your friend would draft. The email contains some
information which is only privy to you two. You will have no inclination to
believe that there is something suspicious about it. This is the level of
sophistication which can be achieved with the help of Artificial Intelligence.
Relatively simple sequence-to-sequence machine learning algorithms
can be installed on an infected device in order to monitor emails of the
compromised victim. It would then study your behaviors and how you respond to
different people. It would also analyze the language and the manner in which
you communicate. Then it can be used to send seriously threatening links and
attachments via email to your friends or people in your contact list. Given
that the email has come from your official email id and in your normal email
writing style, they are bound to open it leading to disastrous results.
These systems will also give personal information which can cause
unintended consequences. A hacker may take advantage of your personal emergency
and use your personal information to steal from your accounts.
Social Media
In today’s time it is hard to find a person who does not have a
social media presence. We tend to put all our details on our websites. The data
present on these websites can be effectively stored in a computer. The computer
can then effectively process this data and can help figure out answers to
security questions. Large scale data breaches are possible just by processing
data which is freely available on the social media platforms. Artificial
Intelligence will help miscreants target a large audience leading to many data
leaks.
Businesses and IoT Devices
Microphones and cameras surround us. We are sprinkling them at every
nook and corner. It had come to light that the CIA had been hacking smart TV
cameras to look at terrorists. If a security agency could do that, so can
professional hackers. A person sitting at some remote corner can have live
footage of us while we are sitting at our homes. Privacy is something that will
go for a toss. Hackers can blackmail people by threatening to release intimate
footage or conversations. Denials of Service attacks are also pretty common
when it comes to IoT devices.
This takes a whole new level when it comes to businesses. There are
readily available codes of speech and facial recognition in the market. An
Artificial Intelligence system can be trained to recognize certain words and
then be used to hit hundreds of business conferencing units. The information
gained from spying on such units can then be sold to the business competitors.
Business secrets, stock market information, etc. everything could be compromised
and taken advantage of in these scenarios.
CryptoCurrency
With cryptocurrencies rising in popularity, cryptocurrency exchanges
are likely to be sitting ducks for attacks, say experts. Security firm FireEye
says it expects to see “more malware stealing cryptocurrency from weakly
protected wallets, shimming password entry to wallets, stealing offline wallets
for brute forcing, or using credentials stolen from the same user”. An AI might
be able to track and identify cryptocurrency users using machine learning. In
machine learning, AIs learn to discern patterns. If an artificial intelligence
were able to identify the behaviors of specific bitcoin owners, it might be
able to determine their identities. AI would be a huge threat to all
cryptocurrency owners if it fell into the hands of criminals.
Data Manipulation
Machine learning promptly relies on data. A machine which is given
incorrect data will produce incorrect results. Hackers instead of attacking
businesses can just manipulate the data and cause much distress to businesses.
Suppose you have submarines collecting data about possible oil
sites. These small submarines do not have a firewall or anti-virus. A hacker
can simply meddle with the data. This can lead to wrong decisions being taken
by the company as a whole. A hacker can practically cause a loss of millions by
simple data manipulation.
This technique is effective across different industries which are
increasing using machine learning in their operations. A tweak in the data they
possess will lead to wrong interpretations and bad business decisions. Hackers
can also demand money from providing original data. The possibilities are
endless.
Conclusion
These are just a few challenges. Besides not needing food and sleep,
computers will be able to make complex calculations much faster than humans. We
will not be able to understand Artificial Intelligence powered attacks in this
earlier phase. But as more and more devices and humanity get connected, the
attacks and their scale shall be evident.
Just as Artificial Intelligence can be used by hackers, it can also
be used by security agencies to detect and prevent threats. Many security firms
are building technologies that will enhance cyber security for businesses and
individuals. The jury is still out as to who will benefit from the advancements;
the good guys or the bad guys?
Comments
Post a Comment