DEV Community

Cover image for How AI is Revolutionizing the Fight Against IP Theft and Insider Threats
Marie Pettit
Marie Pettit

Posted on

How AI is Revolutionizing the Fight Against IP Theft and Insider Threats

It’s important to keep workers happy. Just how much may be determined by the number of Intellectual Property (IP) Theft attempts by disgruntled former employees. However, artificial intelligence (AI) is making catching – and thwarting – these attempts so much easier.

How Intellectual Property Gets Away

According to data security firm Cyberhaven, there are several ways intellectual property theft occurs.

  1. Unauthorized Access | A sophisticated exploit – or a simple stolen password – can grant an unauthorized user illicit access to sensitive information.
  2. Misappropriation of Disclosures | When information gets passed to third parties in the course of business, non-disclosure agreements typically follow. Break those, and you could be guilty of trade secret theft.
  3. Employee Abuse | During times of layoffs, typically benign employees can go rogue. Tempted by outside threat actors looking to prey on their fear and discontent, they can be lured to divulge intellectual property they otherwise would have kept safe. Good old discontent can engender this reaction, as well. Discouragingly, it is reported that 75% of employees have stolen from their employer. In the midst of the digital revolution, how is a trade secret any different from a stapler when it comes to moral dilemmas?

The Sneaky Ways of Insider IP Theft

Intellectual property theft occurs when copyrighted, trademarked, or otherwise proprietary data is illicitly used by anyone other than its rightful owner. Many times, this means sneaking sensitive data out of a protected network via convoluted digital back alleyways. With every new save in every new location, the trail gets a little more lost. Finally, it ends up for sale on a Dark Web forum and no one really knows how it got there. Well, one particular employee might.

The problem comes when you have a network-centric – not data-centric – security stance. More and more, this is becoming outmoded, for several reasons.

Network-centric security protects the location. If that “wall” is hacked, everything inside is up for grabs. Data-centric security, on the other hand, goes wherever the data goes. It often comes in the form of Data Detection and Response (DDR), and this is how it works.

Digging into DRM

Digital Rights Management (DRM) does a lot to change that network-centricity. DRM tools allow you to place protections directly on the intellectual property itself, blocking access to non-authorized recipients, and even dictating what the intended recipient can and cannot do. Don’t want them to copy and paste? No screenshots allowed? Is this View Only? You can add all those specifications in.

What about when your intended recipient gets a layoff notice one week later? Immediately deny access to the files you’ve already sent, even if they were able to access them before. This keeps trade secrets safe in a business deal, allowing you complete protective control until the ink is dry.

DRM allows a sliding scale of trust to be put into place – think of it as the business-side of the Principle of Least Privilege – so that when you send out sensitive data, you always have the upper hand. This not only prevents vital information from falling into the wrong hands, but it prevents potentially nefarious inside actors from carrying out their plans.

How AI is Improving DRM

Now, mix AI with Digital Rights Management and you can do even more to protect intellectual property and the rights of trademark owners, copyright holders, and the corporations and economies they support.

Here’s how.

AI in Image Recognition | AI’s ability to quickly place visual content – out of a swath of multimedia assets across the entirety of the internet – is a boon to DRM tools looking to locate instances of copyrighted material being used without explicit consent. Image Recognition picks up on instances of Watermarking, in which indistinguishable features of ownership have been embedded into the media files, allowing it to catch illegal reproductions. AI capabilities can also extend to automatically initiate take-down requests.

Natural Language Processing (NLP) Delivers Context | DRM can only stop as many instances of copyright abuse as it can find, and the internet has at least 630 billion words (and up to 100 trillion if Reddit is to be believed). NLP provides a way to analyze text data to spot instances of copyright infringement and can also make allowances for more nuanced uses, such as in multiple news outlets simply covering the same topic (as opposed to copying each other’s work).

Catching Up to Criminals

IP Theft is rampant, with criminal arrests for the crime up by 36%. However, as attackers level up, so can we. Thanks to the power of AI-enhanced detection tools, perpetrators of IP Theft can run, but they can’t hide. At least not as well as they used to before. And at the rate AI is developing, they may not even be able to do that for long.

Top comments (0)