A developer’s best friend: Lessons learned from our canine companions about AI code security

Happy International Dog Day! This official holiday celebrates our furry friends and the joy they bring to our lives! Today is particularly special for all of us at Snyk because of our four-legged mascot, Patch the Doberman.

But what exactly does a dog have to do with application security? Here at Snyk, we see the idea of a “guard dog” protecting someone’s home as similar to how AppSec solutions can protect today’s development practices. In fact, that’s why Patch became our mascot in the first place. He represents the idea of a friendly but alert protector, ready to defend his territory from danger. 

As Snyk has acted as this friendly guard dog against AppSec vulnerabilities over the years, we’ve had the unique opportunity to watch the world of application development change dramatically with the introduction of new technologies and larger, more complex supply chain ecosystems. Although much has changed in a short amount of time, our guard dog mentality has remained the same. We prioritize friendliness and readiness to support development teams while fiercely defending their applications against threats. And in the era of generative AI, protecting applications against common first-party vulnerabilities is more critical than ever. AI coding assistants are extremely prevalent in today’s development environments. But because they pull training data from across the web, they can easily perpetuate common security issues.

In honor of this year’s International Dog Day, let’s look at three lessons that we can take from our four-legged friends, especially as related to securing AI-generated code:

1. A loyal companion can change your life for the better.

Those of us who own dogs know how comforting it can be to have a canine friend by your side. It’s reassuring to know that your four-legged friend will sound the alarm if something goes bump in the night, or stay by your side if life’s ups and downs get rough (or should we say, “ruff”?). 

In the same way, developers can build applications with much more confidence when they have an AI security companion. The right security companion can keep an eye on the AI-generated code developers are pushing to the repository and sound the alarm if the code contains security vulnerabilities. That way, the developers don’t have to deal with nearly as many security concerns later down the pipeline; they can find and fix AI-generated code vulnerabilities as soon as they create a pull request. 

2. When your dog matches your lifestyle, you can really enjoy life together. 

Every dog breed has a unique set of characteristics and personality traits. And when you find the perfect dog to match your lifestyle — whether an active breed to take morning runs with you or a laidback, cuddly breed to watch your favorite Netflix shows with you — you can really develop a strong bond with your dog and enjoy every aspect of life together. 

In the same way that each person needs a dog matching their unique situation and lifestyle, each developer needs an AI security companion that works well with their existing day-to-day workflows.

AI security companions that play well with development teams tend to have the following traits:

• Compatibility with each development unit’s favorite integrated development environment (IDE) and CI/CD pipeline tools

• Fast-paced scanning capabilities to keep up with the sheer volume of AI-generated code that tends to enter repositories

• Quick fix recommendations to show developers simple next steps for fixing security issues

3. Training is an important part of owning and caring for a dog. 

Dogs are capable of amazing things when they are trained correctly. But a lot of hard work and repetition goes into getting your dog to behave a certain way. Those of us who have owned puppies understand just how challenging the training process can be.

Because many application security companies (including Snyk) use AI-powered solutions to secure AI-written code, the concept of good training applies here, too. When choosing an AI solution to secure your AI code, it’s essential to find one that has been trained correctly. Otherwise, just like your energetic puppy who hasn’t undergone any training yet, it won’t do what it’s supposed to do and will be more prone to mistakes and accidents (aka hallucinations, inaccuracies, etc.)

Snyk: A security companion for your AI-generated code

Snyk Code, our static application security testing (SAST) tool, can move at the speed of AI with scanning at pull requests and smart fix suggestions powered by our DeepCode AI Engine. And since we leverage AI to fix AI-generated code, we prioritize training our AI on high-quality data. DeepCode combines multiple machine learning methods, human expertise, and security-specific datasets to produce trustworthy results.

Find out more about the importance of AI security companions here.