Security Considerations for the AI Era (ZDNet)


IT pros who are working with machine learning or artificial intelligence deployments cannot overlook the importance of security. AI models can be built with the capability to access billions of data points within a business unit or an entire organization – and many incorporate third party data feeds as well. This type of deep learning may put a company at risk.

What do you need to consider to make AI development more secure?

Secure the code. Code should be designed in a way to prevent unauthorized access. Machine learning is adaptable, so algorithms can be written to reduce risk. AI can also learn the environment and application well enough so that it can detect when changes occur or are inserted.