Artificial Intelligence, Hospitals, Health Tech

Mayo Clinic, Google eyeing ‘AI factory’ as collaboration moves forward

The health system and tech giant are more than a year into their 10-year collaboration. As their partnership evolves, the organizations are looking to create an “AI factory” — an assembly line of AI solutions that are developed at scale and incorporated into clinical workflows.

AI, machine learning, technology

More than a year into their 10-year strategic partnership, Mayo Clinic and Google are looking beyond foundational development.

Their focus will now shift to creating a factory for artificial intelligence solutions, where these technologies are developed at scale and incorporated into clinical workflows, leaders from the organizations said.

This ambition will be aided by Google’s plans to open an office in Rochester, Minnesota, close to Mayo Clinic. This will be the technology giant’s first office in the state, and it will allow Mayo Clinic staff to be in close contact with their Google colleagues.

“As our work moves from building the [technology] foundation to running the [AI] factory, the emphasis will be a lot more on interaction with the clinicians,” said Cris Ross, CIO of Mayo Clinic, in a phone interview. “So, I think it’s really wonderful that Google will have access and be able to walk the halls with some of our clinicians, to meet with them and discuss what we can do together in the medical context.”

Google hopes to open the office later this year.

In the year and a half since Mayo Clinic and Google entered into their partnership, the organizations have been focused on bolstering the health system’s existing technology architecture and moving its massive trove of data into the cloud.

“This foundation will unlock opportunities to implement new solutions to really move quickly,” said Aashima Gupta, director of global healthcare solutions, Google Cloud, in a phone interview. “To move from idea to ideation to deployment.”

One example of the kinds of AI solutions the organizations can develop came last October when they announced a joint AI initiative focused on radiation therapy planning. Mayo Clinic and Google are using de-identified data and images to develop an algorithm that can help clinicians create radiation therapy plans for cancer patients more quickly and efficiently.

Another way in which the organizations are combining their expertise is by using AI to analyze radiology images for body composition markers that can indicate underlying chronic conditions.

But looking ahead, Mayo Clinic and Google want to build an AI solutions factory — an “assembly line of AI breakthroughs,” said Mayo Clinic’s Ross.

“To me, the big picture with our relationship with Google is simply being able to do the kinds of things that people are doing in little bits all over the planet, to be able to do the same kinds of things but at scale and repeatedly,” he said.

The collaboration is looking to move beyond the one-off model that AI development typically operates within. Instead, the organizations will focus on creating repeatable AI models at scale as well as operationalizing machine-learning, said Google Cloud’s Gupta.

“What we are really looking for is the long-term success of infusing AI into complicated clinical workflows,” she said.

To establish this AI factory approach, academic researchers, clinicians and engineers need to be able to come together and engage in a dialogue. The organizations hope that the new Google office in Rochester will help them create this environment, enabling the technology experts to learn from clinical leadership, and vice versa.

Photo: Blue Planet Studio, Getty Images