Meow-lware: What Cats Have to Do with the Future of Cyber Security

Cats have long commanded control of the internet. With their unexpected expressions, wild antics and supreme fluff, cats were some of the first viral memes and continue to dominate on social media. In fact, cats are so popular online that most people believe the urban legend that 15 percent of internet traffic is cat-related. Though the rates of cat traffic are in truth much lower, the millions of kitty pictures and videos on the internet might soon become one of the biggest forces of good.

Infosec firms are desperately looking for tools to help in the ever-escalating arms race of cyber-security, and the latest and greatest strategy is artificial intelligence. There is only one problem: How does one train a machine to recognize dangers it has never before encountered?

That’s where cats come in.

How Artificial Intelligence Works

Movies and TV shows often depict artificial intelligence in the form of androids or humanoid robots, so many people have started to believe that “intelligent” robots have similar thoughts, feelings and experiences as humans. Unfortunately, they don’t – at least not yet. However, they do learn in ways similar to human children -but you won’t see machines making the same adorable mistakes as babies.

First, it’s important to explain what artificial intelligence (AI) is, and how it works. AI is a broad field that encompasses anything to do with imbuing machines with a human being’s reasoning faculties. Machine learning is a category within AI concerned with, as you might expect, helping machines learn.

Just as human children acquire language and skills through constant exposure to information which their brain processes into patterns and insights, machines programmed with certain algorithms and provided with data can gain similar comprehension and ability. And, just as learning is useful to humans (quite the understatement), learning is proving valuable to machines.

Already, AI is becoming an integral element of everyday life: Virtual assistants like the Amazon Echo and Apple’s Siri rely on machine learning as do prediction tools, like you might use for commuting through traffic. Marketing strategies are relying more and more on machine learning to pinpoint individuals’ preferences and provide messages optimized to create and convert leads. Yet, the most important application of all is likely that of cyber security.

The Benefits of AI in Cyber Security

In the last decade, cyber security has escalated almost beyond belief. Today, billions of malware instances float around the web, and dozens more are made every minute. It’s literally impossible for the cyber security tools of the past to keep up – which is why we need a cyber security tool of the future.

The average endpoint protection services of the past started and ended with a relatively rudimentary antivirus product. This software was signature-based, meaning it contained a catalogue of known malware signatures, or strings of code, and attempted to match files with any in its library to identify threats. This was a time-consuming process, demanding excessive processing power and frequent updates to keep up with the multitudes of new viruses and worms emerging daily.

Equipped with AI, cyber security becomes more efficient. Intelligent cyber security tools can predict malware before it emerges, protecting devices and networks more thoroughly than ever before. It’s likely that enterprise security systems of the near future will not be complete without some type of AI tool, and from there AI security should move into the consumer market.

But, wait, weren’t we talking about cats?

Why Cats Are the Answer

Though AI technology is promising, it is hardly a well-developed product. Security firms working to develop AI tools are finding it difficult to train their programs using the malware data they have available because much of it is out of date. Unfortunately, failing to use the most current malware strains could cause programs to develop patterns that aren’t effective at recognizing future malware.

That’s where cats come in. Cats come in all shapes, sizes and colors, but humans can always identify a cat from a non-cat. Because there are so many pictures of cats online – black cats, striped cats, fat cats, floofy cats, hairless cats, deformed cats, etc. -AI has abundant opportunities to learn what a cat looks like, and developers have no problem creating AI that can recognize any type of cat immediately.

Thus, developers realized that they need to find a way to simulate malware in all its diverse forms without actually subjecting their budding security tools to infection. Fortunately, a cyber security firm named Endgame successfully compiled more than a million representations of malicious files – and it made the data set open, so anyone can use it to train their machines.

Once again, thanks to the inspiration of cats, the world is a better place.