My Artificial Intelligence is Better than Your Artificial Intelligence – Or How to Truly Stop Malware
- At RSA 2017, there was one resounding message that attendees were bombarded with; artificial intelligence is here to solve our security problems.
- The problem is, cybercrime continues because the bad guys are using the same techniques.
- It’s time to change the game.
For years, it’s been humans vs. humans. Cybercriminals and cybersecurity organizations constantly change TTPs in an attempt to outsmart each other. The introduction of AI to solve for cybersecurity helps exceed the human capacity to respond to the ever changing TTPs of cybercriminals. But, it’s no silver bullet.Cybercriminals also use AI.
The age old game of cat and mouse, or cops and robbers is still going strong in the cyber industry. That’s because there are three key principals that are not changing anytime soon:
- There will always be software vulnerabilities
- Malicious actors will always take advantage of vulnerabilities
- You cannot anticipate the adversaries next move
It’s sad to see an industry where it seems vendors are chasing their tales with marginal advances in bolt-on security designed to detect and stop threats. Don’t get me wrong, AI does help improve security, but the cybersecurity industry needs to move beyond a detection strategy to a cyber resilience strategy where security is built-in by design.
Even though AI provides better “mouse traps”, breach rates are still increasing and cybercriminals are making over $3 trillion in profit. Polymorphic malware changes in less than 60 seconds to something else where detection-based solutions are becoming less effective.
At the RSAC Cryptographers’ panel, Whitfield Diffie calls for a change the way security is delivered today using the hardware-enforced security to isolate threats. He goes on to explain that we can have vastly more secure systems using hardware-based security.
To further emphasize the need for a change in the way we deliver security, at RSA this year we demonstrated how easily AI can be bypassed with polymorphic malware with a simple repack example. If you missed it, here is a short demo video.
It’s time to look beyond AI and failed detect-to-protect model and look to something truly transformative – virtualization-based security. Join our customers who arehaving success.