Software

5 Questions to Ask About Buying AI-Enabled Security Software

Before you make that purchase, check to see that a new product contains actual AI services and isn’t just hype.

Mike Chapple is associate teaching professor of IT, analytics and operations at the University of Notre Dame.

Security products incorporating artificial intelligence techniques may reduce the workload for human analysts, taking over the time-consuming job of correlating information sources and mining voluminous logs to uncover suspicious patterns of activity. Vendors, seeing the hype around AI, are quick to slap the label on almost any technology for a cutting-edge veneer. Here are five questions to ask before purchasing an AI-enabled security system:

1. What Data Was Used to Train the AI?

AI and machine learning techniques detect patterns in data and use those patterns to make predictions about the future. But those models are only as good as the data used to train them. Make sure you understand how the security system’s models were created. Does the product create a model customized for your organization, or does it use the same model across all types of environments?

2. Does the Technology Truly Use Machine Learning or AI?

Some vendors assume that if their product is making recommendations, it’s using artificial intelligence. In reality, some of those products are simply using the same pattern-matching rules that they’ve always used, rather than relying upon more sophisticated data models. If you don’t understand the answers you receive, that’s a red flag that should prompt some hard questions.

3. How Does the AI Learn and Improve Its Abilities Over Time?

As networks, systems and user behavior change, models must adapt to reflect those changes. How is the model updated? Is it continuously retrained on your new data as it’s collected, or does it require product updates from the vendor to improve its abilities? How are those updates delivered?

4. Who Owns the Resulting Models?

If the product is using your data to either develop or contribute to broader models, do you have any ownership rights to those models? Is the vendor permitted to use that model with other customers? Are you able to access details of the model and use them in other security applications? Read the contract’s fine print to make sure you clearly understand how the vendor will use your data and what rights you retain.

5. Will the AI Save You Time or Improve Your Security Posture?

You should invest in AI technology only if you truly believe that it will either reduce the amount of time currently spent by security analysts on your team or provide you with insight that you don’t receive from your existing products. As with any business investment, make sure that the expected ROI justifies the cost.