Cybersecurity has come a long way in the last twenty years.

In the 90s, the predominant security model used to create secure operating systems was the “castle and moat” approach. Everything inside the firewall was trusted and anything outside it wasn’t trusted.

But emerging internet services like email meant that things needed to get through the wall. This was the beginning of the antivirus era of cybersecurity, an era that we are still in. Antivirus works by identifying a threat, creating a signature, and distributing that signature so that every other computer with antivirus software installed can identify malware and defend against it.

See also: Is cybersecurity for smart cities being dangerously underestimated?

Though the cybersecurity model hasn’t changed much since the advent of antivirus software, that could be about to change, thanks to advancements in cyberthreats.

Most people creating malware use it once, and never again, which means identifying it and protecting against it in the future isn’t as helpful as it once was. A lot malware is advanced enough to slip through signature-based techniques of identifying them. Finally, the sheer volume of cyber threats continues to grow at an exponential rate and it’s getting harder to stay on top of them.

Deep learning and the future of cybersecurity

Advancement in the field of deep learning allows artificial intelligence developers to create machines that can think like humans but process vast amounts of data quickly. Artificial intelligence researches are hopeful that AI may be the answer to the growing cyber threat problem. AI could theoretically identify eliminate cyber-threats as fast as they can be created.

While previous methods for protecting against cyber threats has been reactionary, the malware attacks, the antivirus software identifies it, and then makes other computers immune to it, cybersecurity led by AI could take a more proactive approach in dealing with cyber threats.