AI holds a “promising future” for the cyber-security industry
As artificial intelligence (AI) continues to disrupt every part of our daily life, regional experts expect the technology to bring with it a number of solutions for the cyber-security industry.
From identifying vulnerabilities in the cyber-space in a faster manner to aiding authenticity detection, AI is said to hold a promising future for the cyber-world.
“AI enables intelligence extraction from massive amounts of data, which would not be possible by humans,” said Michail Maniatakos, Assistant Professor of Electrical and Computer Engineering at New York University Abu Dhabi. “At the same time, it can learn from fuzzy, noisy and incomplete data, which makes it different from older generation data science and automated scripting.”
“AI enables intelligence extraction from massive amounts of data, which would not be possible by humans”
Threat-wise, AI can aid in intelligence extraction, such as understanding network layout, and automate malware generation, in relation to defense. It can also understand more sophisticated malware and viruses, and infer attack patterns.
For organizations and their efforts towards ensuring password protection, AI can help differentiate between human and script behavior. “In more detail, you may have forgotten your password and try to recover it, or a computer is trying to recover your password,” Maniatakos explained. “Humans and computers would have different behaviors and AI can be used to pick that up.”
Authenticity detection is considered a massive domain, but the principles are said to remain the same, whereby AI can learn very distinct characteristics, not easily detected by a human, to aid this authenticity detection. In addition, AI can easily scale, including checking the authenticity of 1,000 objects per second. “Humans have fundamental biological limitations on what they can do,” he said. “AI can scale very easily – just add more computers to it. At the same time, the theory behind AI, as well as its computational efficiency, are also rapidly improving. My personal opinion is that, today, AI is developing much faster than hackers.”
As the AI-world heavily depends on data, many questions on privacy arise. The answers, he believes, are application-dependent. “For example, for self-driving cars to recognize street signs, the data used to train the model can be generic without personal data,” Maniatakos added. “Generally, it is always a good idea to anonymize data before feeding them to any AI model. But vulnerabilities at hardware and software level will always be there – AI can assist in identifying them faster, but it cannot, as of today, eliminate them.”
“The theory behind AI, as well as its computational efficiency, are also rapidly improving”
According to Matthew Cochran, CEO and Co-Founder of URS Laboratories Ltd in the UAE, AI is not the future, rather, it is here now. “AI helps us complete tasks faster, cheaper and without the same mistakes as a human user at a scale unseen in our past,” he explained. “It also provides this same computer power to those that might attack us via cyber-hacking and the same tools we see as a solution in the wrong hands can be a threat 10 times greater due to the lack of defensive AI on legacy systems.”
He said the time for organizations to prepare is now before they are hit by an attack on their AI platforms or Big Data. Getting experts hired that understand AI on their security and IT teams while building redundancy into their systems and back-ups will prove key in protecting themselves in today’s world. “AI actually makes passwords a thing of the past,” Cochran noted. “Most passwords follow rules and thereby can easily be determined in a matter of minutes by AI hacking password tools. On the same hand, AI can randomize the numbers, letters and symbols in traditional passwords to make them safer but, in the future, biometrics in my opinion will replace that, so AI does not and cannot currently replicate human biometrics.”
“AI can randomize the numbers, letters and symbols in traditional passwords to make them safer but, in the future, biometrics in my opinion will replace that, so AI does not and cannot currently replicate human biometrics.”
According to a study by Senseon Pro, almost 70 percent of small and medium enterprises are planning on implementing AI in the next five years while 81 percent believe AI will be fundamental to the future of cyber-security.
“AI is basically our new electricity,” said Dr Fadi Aloul, Professor and Head of Computer Science and Engineering at the American University of Sharjah (AUS). “It is something we cannot live without and the UAE is one of the leaders worldwide in adopting AI, which is why we have a full ministry dedicated to it.”
With the number of sensors, such as IoT devices, that stream Big Data rapidly increasing in number, smarter results are inevitable thanks to AI in the long-run. And the adoption is timely, as a study by the World Economic Forum last year revealed that cyber-attacks are the third most likely risk we face today, following extreme weather and natural disasters. “It could destroy a whole government, which is why it is important for us to come up with new techniques to improve it,” he explained. “With hundreds of emails and data received every day, AI can be utilized to flag whether an email is spam or a phishing email that contains viruses or links to malicious websites. In the past, many of those would go through the filter.”
As many companies that have experienced attacks in the past were initially hit by a malicious email sent to a non-technical member of staff, Dr Aloul spoke of a greater control over the threat today. “That is the first intrusion into a company,” he said. “But AI can play an important role in that, as well as capturing if there is any data being leaked from the organization, any viruses being downloaded, or any passwords that are easy to guess – AI can come up with intelligent combinations.”
According to PricewaterhouseCoopers (PwC), AI will become an important part of every major organization’s cyber security’s toolkit. “Scalable machine learning techniques combined with cloud technology are analyzing enormous amount of data and powering real-time threat detection and analysis,” it said in a report on the Global State of Information Security last year. “AI capabilities can also quickly identify ‘hot spots’ where cyber-attacks are surging and provide cyber security intelligence reports.”
Dr Aloul described the new AI world as a “cat and mouse game”, in terms of keeping up with the technology for the greater good. “In cyber-security, AI is a must, given that we know that many of the attacks today are cyber-related, and not physical,” he concluded. “We say, today, that people have to think through security for anything that they are doing. It is something you develop and you learn as you grow, but AI will definitely change the world when it comes to cyber security”.
“Scalable machine learning techniques combined with cloud technology are analysing enormous amount of data and powering real-time threat detection and analysis”
The International Exhibition for National Security and Resilience (ISNR) Abu Dhabi is set to take place in Abu Dhabi from March 17 to 19, 2020, with AI as its main theme.
Co-organized with the UAE Ministry of Interior, ISNR Abu Dhabi will bring the national and cyber security community together to accelerate public-private collaboration for a safer connected world.
Four dedicated communities on National Security, Cyber Security, Policing and Law Enforcement and Critical Infrastructure Protection will deep-dive into challenges and explore new technologies and best practices to avert the ever-growing threats they are facing.