Google DeepMind exec on AI rising effectivity

Employees in Google DeepMind’s legal department might have found an answer to AI’s existential query. Google DeepMind is now using its Gemini AI model to handle roughly 40% to 50% of information requests sent to its legal department, Google DeepMind director of strategic initiatives Terra Terwilliger said at Fortune’s COO Summit in Middleburg, Va. on Monday. Google DeepMind is an AI research lab and subsidiary of Google.

The legal department at Google DeepMind uses AI for “finding information that you know you have somewhere, but you can’t put your hand on it right then,” Terwilliger said. While it’s “the basic use case of simply retrieving and finding that information,” she said, “it saves a remarkable amount of time.” 

Google isn’t the only company using AI for rote tasks involving paperwork. Bayer’s chief operating officer of its global pharmaceuticals division, Sebastian Guth, said the company regularly uses AI to fill out parts of the reams of regulatory filings it is required to compile as part of the research and development process for its medications. Guth estimated the company used AI to fill out between 70% and 80% of these dossiers required by regulators around the world. 

While the particulars of these two examples might be different, both Bayer and Google principally use AI to make their organizations run more smoothly. Guth termed these “workflow efficiencies” during his on-stage interview. 

“There is plenty of paperwork to be automated so that high-skill people can go do more high-value tasks,” Terwilliger added. 

Implementing AI into work flows

Getting employees up-to-speed on new AI use cases won’t happen overnight. That means companies will have to make an effort to train employees on how to best use AI, even in more straightforward instances. Terwilliger urged the audience of COOs to remember that training employees to use AI will require time and effort. 

COOs should “make sure we’re valuing the time that people are putting in to learn how to use these systems and rewarding them for doing so,” she said. 

Guth had his own word of caution. “In large organizations, there’s at times a risk to using technology for technology’s sake,” he said. “In my mind, at the end of the day, it’s a means to an end.” 

The end goal for Bayer is to develop drugs and medicines faster than had previously been possible, he added. “It’s not about chasing shiny toys just for the sake of chasing them,” Guth said. 

However, these early forays into AI won’t mean the technology will slowly replace scientists and researchers that work in the pharmaceutical industry, according to Guth. 

“I actually don’t think that AI is going to steal jobs, because, hey, unpacking 3 billion years of evolution in a cell is pretty damn difficult,” he said. “It will continue to require human capacity and the art of science.”