A major healthcare provider, one of the largest insurers in the country, and an educational robotics company all touted artificial intelligence as the future of work at Hartford’s inaugural AI Day March 9.
The event—run by LAUNC[H], an organization dedicated to helping Hartford innovators and entrepreneurs—showcased area businesses leading the AI space.
Three CBIA members—Hartford HealthCare, Travelers, and Movia Robotics—did a deep dive into how AI is transforming the future of their industry.
Dr. Barry Stein, vice president and chief clinical innovation officer at Hartford HealthCare, spoke about AI's ability to “transform healthcare.”
He began by acknowledging the competitive advantage for embracing the technology.
“There’s a lot of friction in healthcare,” Stein said, “but all that friction is an opportunity for innovation.
“There are thousands of startups out there that are attempting to take a piece of the healthcare ecosystem, and almost every one of these companies has AI driving their solutions.”
In other words, if Hartford HealthCare does not embrace AI, a competitor will.
The two major categories of innovation AI has driven at Hartford HealthCare are clinical and operational.
With clinical predictive analytics, the company can make predictions about treatments and likely diagnoses of patients in advance.
For example, it can forecast which patients going through chemotherapy will have to be readmitted and which patients will develop strokes.
Using AI to make these predictions helps Hartford HealthCare prepare treatment for patients ahead of time, Stein said.
Operational analytics helps the company in logistical preparation.
When COVID-19 began, Stein said businesses struggled so much because nobody was prepared for what was to come.
“It was literally like flying in the dark,” Stein said, adding, “how do you operate a massive healthcare system when you have no idea what to expect?”
But using operational predictive analytics, Hartford HealthCare was able to adequately prepare for a massive surge of patients.
AI helped the company predict how many patients there would be, how many beds were necessary, and how much PPE to order.
“This saved many lives,” Stein said,” and this is the kind of thing that is so impactful, and we could not have done it without AI.
“The opportunities to solve problems are endless,” Stein said, “and healthcare is just an incredibly exciting domain to be in.”
Two representatives from insurance company Travelers spoke about the benefit of AI in their underwriting program.
George Lee, senior director of A.I. accelerator, and Murat Yasar, A.I. research lead, both see the technology as the next step in making their work more accurate and efficient.
The two began with an example of the important yet time-consuming role underwriters play in determining roof conditions for property insurance.
In some cases, an underwriter could have to look at 100 different roofs using online mapping software to determine the condition—a laborious process that cannot be done quickly.
Using AI, Travelers created an image segmentation classification model that could scan the images and determine the quality of roofs.
This process, which normally takes months without AI, can now be completed in a matter of hours.
“We can make our underwriters much more efficient, as well as make our customer experience much better,” Yasar said.
This software can also be used to assess damage from natural disasters such as wildfires and hurricanes, Lee said.
“After these events occur, it’s a logistical nightmare,” he said, as roads are closed, power is down, and it is extremely difficult to send underwriters out to assess the damage.
But using “best in class, state-of-the-art models” makes the process much easier for both the company and their customers.
Tim Gifford’s company Movia Robotics uses robots to communicate with and teach children on the autism spectrum.
Gifford’s research found the robots create a simple and nonjudgemental environment to teach children who typically have trouble communicating.
The robots are not fully autonomous yet as the company is relatively young, so there is always an adult present to facilitate the interactions with the children.
Much of the interactions between the robots and the children comes through what Gifford calls “contextual AI.”
“Sometimes the robot leads the child, and sometimes the child leads the robot,” Gifford explained.
“It’s not just the robot saying ‘follow me, do it,’ sometimes the robot is seeing where the child is and is meeting the child there.”
The company’s AI software allows the robot to fluctuate between different modes of interaction, such as being a teacher, a coach, a therapist, a friend, and even a playmate.
“To the child, the robot can do all these things. It’s like a dynamic, fully formed person.”
Gifford said he collects data from interactions between children and the robots in order to keep improving his product.
Movia Robotics wants to use the collected data “to drive settings for future experiences and to modify the behavior of the robot as they’re interacting with a child to better optimize their interaction.”