What is the Next Cybersecurity Crisis?

Let me give you a hint. It is not what you might expect.

It is not social engineering.

It is not hackers breaking into your network.

In fact, it is not any of the “usual suspects”.

It is about whether you can TRUST the data you are looking at.

In one sense, it is about AI. We are using AI for everything and if the AI’s training data is off by a little bit, all bets are off about the validity of the data you are looking at.

And even if you are not using AI directly, you are receiving data from all over. Maybe the web, maybe from customers, maybe from vendors.

How do you know if you can trust it?

Say you have some sales data. Okay, if it is YOUR sales data you probably know if you sold something or not, so probably that is solid. But then you want to “enrich” that data with data you bought from a third party. What about data quality then?

I will share a secret with you. Actually, it is not really a secret since I have told a handful of people about this.

You know those forms that you fill out on the web when you just want to get a little information about a company or product. I know that the company would really like that information so that they can bug you. Sometimes they just tick me off. So I make up information. I lie about my age. My address. Even my phone number. If enough people do that … I remember one person I told was amazed. He thought the data they were getting was clean.

Now what if your AI personal assistant is doing research for you. Does your AI tell the truth? Maybe. Maybe not.

And companies are now using AI to create content for their web sites and blogs (NO! I AM NOT AN AI. 🙂 ) Is that data accurate.

Then the AI hoovers up that data that another AI created. It spreads the disease like someone with measles going to a concert.

Sometimes you MAY be able to detect bad data. If the data says that I am 150 years old, it is probably not accurate.

But what it says I am 60. 50. 40. 30. That one might be cross-checkable. Depending if the name or the address are real.

But not all data is easily verifiable.

We often have questions about the number of visitors that Google says watch our YouTube videos. That is much harder to check. Maybe you can insert beacons in it, assuming Google doesn’t block them from working.

But you get the idea. You can no longer assume your data is real.

If this raises more questions than answers, give us a call.

Credit: Security Week

by