Securing Artificial Intelligence

Google , AWS and Azure all are heavily investing in AI and enhancing their AI capability at a faster pace. So, as security personal how do you keep up with it?

Today’s blog helps you to look Artificial intelligence from Application Security eyes.

In Modern App architecture, i would include Infrastructure to the Application code . So it is imperative to secure both.

Application needs to be secured at different phases

• Securing during coding phase
• Securing code during the Pipeline
• Securing application during the runtime.

Just like application, AI needs to be secured at every phase . But we need to understand how AI phases actually look

Snip from OWASP AI Security and Privacy.

• In Al, it all starts with Data. With the Initial set of data you define what data is required for your use case, you get it using the DataPrep Code.
• Now , with DataPrep Code you have the data in the format of inputs and output that would you train the Algorithm. So, now you need a code to train algorithm . That’s where Train/Test code comes in.
• Next, you would need to add the algorithm code to your application which , Now algorithm becomes comes part of your application. Eventually your Application code

So, Its is Important to look AI Security in the eyes Application Security. But isn’t enough , Probably NOT ! we look into securing AI more in next blog!