- Benjamin Kang, Professor Yu
Are organizations truly secure with the rapid developments in technology?
Our thoughts on Cybersecurity for the post-quantum age.
Given the rapid rate of technological advancement, we have observed requirements where algorithm processing performance is a key requirement for artificial intelligence solutions in the current age. Gone were the times where algorithms needed hours to train a model with the rise of tensor processing units (TPUs), organizations can train their deep-learning algorithms in a matter of minutes.
What does these exciting-yet-scary developments entail for Post-Quantum Cybersecurity? Our perspective for post-quantum cybersecurity are embedded deeply in our product design principles, namely:
1. Authenticate First, Connect After – Software defined perimeters to provide ‘invisibility cloaks’ against external attacks
2. Security High Wall – Segregate networks to mitigate damage of attacks.
3. Zero Trust Policy – Trust nothing, sanitize every file.
4. AI Cybersecurity tools – Use AI to augment defences and analyse attacks (anomaly, threat detection, explainability, robustness)
Based on these principles, we have developed the world’s first driverless data diode and a cloud variant with our cloud provider partners to secure enterprise critical infrastructure. In conjunction with our network segregation products, we are developing tools to explain, determine and enhance robustness of artificial intelligence algorithms.
In conclusion, we believe that Post-Quantum Cybersecurity will require organizations to proactively explore new approaches to secure their critical infrastructure. Organizations should consider evaluating robustness and explainability of their artificial intelligence algorithms as malicious cyberattacks can occur from external parties or malicious artificial intelligence algorithms.