At JFrog SwampUp 2023, the buzz with all about AI, especially with JFrog’s announcement of Machine Learning (ML) Model Management capabilities. These conversations around AI and ML reflected these technologies’ growing influence and importance in the DevOps world. How much of the generative AI conversation is just hype, though? And what does that mean for AI and ML as a whole? Alan Shimel, CEO of Techstrong Group, and I sat down with Stephen Chin, VP of DevRel at JFrog, to find out.

As far as Chin is concerned, even as more companies create and leverage AI models, these models must be managed like any other software component. Chin said JFrog Artifactory acts as a staging ground to operationalize models using DevSecOps best practices. Algorithms and models will continue to grow in size and complexity, and they will require robust processes around deployment and management – just like any other software artifact. The key, Chin said, is to think of ML as just another development language and leverage tools that standardize and streamline working with it. Compared to traditional enterprise applications, though, DevOps workflows for AI/ML are still relatively immature, but Chin said JFrog’s new model management capabilities aim to provide that missing automation and governance using DevSecOps best practices for governance, security, and licensing compliance.

Additionally, Chin noted that AI/ML have become essential for development teams to keep up with the explosive demand for code. At this point, AI has become table stakes. In the AI arms race, the winners are those who understand AI has become a vital development tool to enhance productivity. In terms of job security, the losers are the ones who can’t keep up with the volume of code. According to Chin, you are out of the running if you don’t embrace AI.

Looking ahead, AI will not make developers obsolete, though – quite the opposite. Given the quasi-unlimited appetite for new code, Chin emphasized that developers who embrace AI will have more work than ever. One way to think of it is that AI provides a new form of “outsourcing” to boost human productivity, much like previous waves of innovation in computer science.

When it comes to security, there are still challenges that need to be addressed; code generated by today’s AI solutions still has significant drawbacks like potential data bias, lack of explainability and simple errors. In the long term, though, Chin believes AI itself will provide the solution to secure an exponentially growing codebase, given its superior scale. Just as AI will make individual developers more productive, it can also supercharge security teams – but it can also empower bad actors. The key will be continuing to democratize the benefits to even the playing field. AI promises to be a transformative technology on the scale of the Bronze Age or Quantum computing, Chin said, but the path forward will require humans and machines working together to ensure it’s used for good.

It’s clear that the pace of innovation in AI and ML is rapidly accelerating. As these technologies become further democratized and integrated into developer workflows, they promise to transform how software is built and secured, Chin said. Companies must take advantage of this technology innovation by providing the pipelines and governance for this software revolution, he added. Chin believes the future will likely see a more symbiotic relationship between humans and AI assistants collaborating to generate, refine and safeguard code. By proactively shaping the responsible use of these technologies, it’s possible to maximize the benefits while minimizing the risks. While this journey has only just begun, the destination is an exciting new era of human potential unlocked by AI, Chin said.