For decades, the police and drivers with lead feet have engaged in a war of radars and radar detectors. Every time police radar technology improves, so do radar detectors to outsmart it. The same is true with cybersecurity. Firewalls are strengthened. In turn, the technology that hackers use is also improved. It’s a never-ending cycle with criminals often gaining the upper hand.
It makes little sense to continue this way. That’s why many cybersecurity experts are focusing less on perimeter security and more on detecting behavior within systems that raises red flags.
AI may be the best way to accomplish this.
Understanding AI Software Security
The problem with traditional antivirus software and firewalls is that once they are breached, it can takes weeks or months for anyone to realize anything is amiss. In the meantime, data is being quietly collected and stored to be sold on the dark web or used for other purposes. AI security software works within systems, learning how each piece of software works, how data is accessed and stored and what user behaviors are the norm.
The more AI learns about the system, the better able it is to create algorithms to determine when a violation has taken place. It can then act to isolate the harm done and inform security personnel of the existence and nature of the cyberattack. Unlike other cybersecurity methods that involve rules based systems, thanks to machine learning, AI solutions can continually learn and adapt. This is in contrast to other methods where rules are added only in response to attacks that have already occured.
Industries Where AI Security Will Be Applied First
While AI can provide security benefits in all sectors, including home and personal use, there are several industries whose cybersecurity needs are more pressing. These are likely to be the industries in which these solutions become mainstream first.
The first industry that comes in mind is health care. The Health Insurance Portability and Accountability Act (HIPAA) and other regulations mean that it is extraordinarily important to maintain privacy protections. Not only is it important to preserve robust security protocols to stay within regulatory compliance, but it can be a matter of life and death. Cyberattacks are becoming much quieter and more gradual. Where previously a cyberattack might steal, delete or alter multitudes of records, leaving an obvious footprint, now attacks focus on watching data over time and making gradual adjustments. This can make changes to patient records difficult to detect. As a result, wrong treatment protocols could be needed.
Academia is another area of concern. Student data has much potential for exploitation. All a hacker needs to do is access a college database, gain information from a company offering tutoring services or providing information on ratings of college writers or retrieve student housing records. This information can be used for identity theft and other nefarious purposes.
Financial services is another industry that comes to mind. J.P. Morgan, Equifax, Deloitte, Card Systems Solutions, Citigroup, the Federal Reserve Bank of Cleveland and several other financial services companies have one thing in common: they have been hacked and, in some cases, multiple times. It should be clear at this point that finance is an industry for which something more robust is needed. AI could provide that solution.
We’ve all been there. A story comes across Facebook or your cable news provider. Target or some other retailer has been hacked. If you’ve shopped there in the last few months, your account information may be at risk. Not only are these attacks bad for a retailer’s reputation, but they can also leave them liable for damages. This is another area in which AI could be ideal.
Finally, it’s becoming clear that local, state and national governments are at risk as well. The commission on elections, the Democratic National Committee, multiple state and local governments and even the U.S. Department of Homeland Security have been compromised in cybersecurity incidents. These attacks threaten national security and can impact elections. AI security can provide ways for these agencies to react to attacks in real time.
How Hackers Will Respond
Just like IT security companies are beginning to explore the use of AI over other methods, hackers will predictably do the same. In fact, there have already been instances in which hackers have employed AI to infiltrate networks. Many of these efforts involve using machine learning to gain as much information as possible about the systems involved. The more that is learned about these systems, the better these bad actors can become at covering their efforts.
Basically, AI teaches malware how to behave in a way that appears perfectly normal. So far, the efforts have not been very sophisticated. However, AI as a cybersecurity solution is still in its infancy. Because of this, it’s difficult to predict if this will become the same cat and mouse game that has plagued perimeter-focused IT security.
Commonly used cybersecurity methods simply aren’t working. Hackers are becoming too sophisticated and predicting future attacks based on past attacks simply is not working. While it’s still in its proving stages, it looks as if AI may just be able to cover the ground where traditional IT security efforts have failed. It will be interesting to see where these solutions will be most effective.
About the Author
Luisa Brenton is a writer in a variety of venues – academic, business and online marketing content. She is a frequent contributor to topwritersreview.com, a review website that evaluates custom online writing services. Find more on Facebook and Twitter.