When thinking about artificial intelligence, the old phrase, “You are what you eat,” comes to mind, and what AI eats is data.

Understanding where the data comes from and how it will impact the model is critical, according to Graeme Thompson (pictured, right), chief information officer of Informatica Inc.

“For the companies that have done the hard work to understand the lineage of the data through a catalog to understand how the data’s accessed using a data access management solution, and then to make sure that the data is of high-quality — those companies are going to go way faster and ultimately end up ahead of the companies that can’t do that,” he said.

Thompson spoke with theCUBE’s Rebecca Knight (pictured, middle) and Rob Strechay (left) at Informatica World, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed what companies are facing with the influx of AI models and how they need to prepare their data. (* Disclosure below.)

Balancing risk prevention and innovation with AI

Without doing the necessary housekeeping for their data storage, companies are going to struggle with safely implementing AI models, according to Thompson. ChatGPT was the software that launched AI into the public forum, but he emphasized that tuning large language models for enterprise data is a very different process.

“We have to caution ourselves and think that, OK, ChatGPT is built on top of the internet. You could argue that the internet is a reasonable proxy for the general knowledge of the world,” Thompson said. “But if you reflect on how that is going to translate to an enterprise use case, my data’s not on the internet. Google hasn’t been organizing it for 20 years. Most company’s enterprise data is a mess. It is as far from ChatGPT on the internet as you can possibly imagine.”

Despite the risk inherent in feeding AI your data, Thompson believes in striving toward innovation. He highlighted Informatica’s Center of Excellence, adding that having legal and security teams present will benefit all team members involved.

“We have to do the hard work to build the trusted data platform, so that when the tech is ready, we are an enabler and not an inhibitor,” Thompson said. “CIOs and data people often end up having to choose between being an enabler or being in the way, and this is one of those moments.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE Research’s coverage of Informatica World: 

(* Disclosure: Informatica Inc. sponsored this segment of theCUBE. Neither Informatica nor other sponsors have editorial control over content on theCUBE or SiliconANGLE.)

Photo: SiliconANGLE