Cyber-security has gained increased attention lately as we see systems being breached for blackmailing, sabotage, and industrial espionage.
Deployment of industrial AI may make the system more vulnerable, and reduce the potential value of the deployed AI system
Cyber-security cannot be achieved through a simple algorithmic “plug”. Instead, it is the sum cryptography, hardware protection, detection, and processes that need to be updated as the both the system and the threat environment evolve
Cyber-security adds cost to an AI solution; It may restrict the impact of the envisioned deployment
There is generally lack of understanding of its importance and a resistance for implementation as “just another IT hurdle”
ROI is hard to quantify
SOLUTIONS ATTENDEES MIGHT WANT TO KNOW
How does one go about considering cyber-security?
What are the different phases of cyber-security?
When does cyber-security need to be considered;?
What do you do when there is a breach?
TAKE AWAY FROM THIS TOPIC
Cyber-security cannot be ignored when we deploy AI systems
Integrate cyber-security into AI solutions as a matter of design principle
Cyber-security threats may arise at the periphery of a deployed system, i.e., not in the core algorithms but in the deployed sensors, the edge devices, the network, the maintenance system, etc.