Artificial intelligence matured with the introduction of generative AI through tools such as ChatGPT and Google Bard.

But as exciting as the possibilities are, new-age AI still isn’t perfect. There’s a need for large-scale monitoring and evaluation of the underlying large language models — especially in situations where mistakes can be extremely costly. That’s where DataRobot Inc. comes in.

“I think the truth of the matter is there are 140 or more open-source LLMs,” said Ted Kwartler (pictured, left), field chief technology officer of DataRobot. “But you need to have an ability to evaluate them. And I think the monitoring and governance aspect is where a place like DataRobot really shines. We can help measure for toxicity, cost, truthfulness [and] a bunch of different dimensions — all on one platform, no matter which LLM you choose.”

Kwartler and Arvind Thinagarajan (right), head of data science and analytics at Gannett Co. Inc., spoke with theCUBE industry analysts Lisa Martin and Dustin Kirkland at the Google Cloud Next event, during an exclusive broadcast on theCUBE, SiliconANGLE Media’s livestreaming studio. They discussed how the Gannett use case exemplifies DataRobot’s AI value proposition. (* Disclosure below.)

Gannett leveraging the platform to favorable results

Gannett is a mass media holding company with names such as the USA Today Network under its umbrella. While gen AI presents vast opportunities in the media space, the company sought to be measured on how it leveraged the technology, according to Thinagarajan.

“At Gannett, we want to be really thoughtful and measured about how we leverage generative AI,” he said. “From an application standpoint, we are prioritizing use cases that don’t impede the way we want to integrate the nuances of generative AI in the end-to-end workflow of an application.”

In a nutshell, the company is embracing AI, but in such a way as to maintain certain guardrails and preserve that human touch in discernment and ingenuity — and it’s tapped DataRobot to that end.

“Having the ability to edit it or approve it and send the feedback loop [on] whether the application works well or not,” he explained. “And then there are so many steps in the process, from getting the right LLM on board [to] development, maintenance of the model, monitoring the model, [and] adding the governance layer.”

Here’s the complete video interview, part of SiliconANGLE’s and theCUBE’s coverage of the Google Cloud Next event:

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

Photo: SiliconANGLE