Like many these days, Severence MacLaughlin foresees a big, bright future for artificial intelligence.
Unlike many, the Rhode Island-based data scientist was willing to quit his job, sell his house and cash in his investments so he could raise money and focus on launching a new AI company in 2019, a high-stakes bet that he could develop groundbreaking, data-crunching technology in health care.
“I went all in,” MacLaughlin said.
More than five years later, his North Kingstown company, which now also has offices in Palm Beach, Fla., is using immense amounts of health data and predictive algorithms powered by machine learning to grasp a patient’s condition and offer predictions
of future health problems in areas such as renal disease, diabetes, cardiovascular disease, behavioral health and respiratory disease.
The idea: flagging people for early health interventions could save lives and millions of dollars in future costs.
Now the company – DeLorean Artificial Intelligence, a reference to the time-traveling car in “Back to the Future” – has signed on hundreds of clinics and several insurance companies, including the giant UnitedHealthcare Inc.
MacLaughlin is not the only one who has been betting on AI’s increasing role in health care, and local advisers in the field say Rhode Island in particular – with its mix of first-rate higher education institutions and health systems in close proximity – is proving to be a good place to take a chance. And they say health care might be the area where AI can really be revolutionary.
Aidan Petrie, managing partner at the Providence-based New England Medical Innovation Center, says there are dozens of medical-technology startups in the local pipeline, many of them employing artificial intelligence such as machine learning to perform tasks faster and more accurately than humans.
And then there are the schools such as Brown University churning out students with big ideas, along with their professors and other independent entrepreneurs across the close confines of the Ocean State.
“Every piece that you need is available and accessible and actually incredibly collaborative,” Petrie said. “That does not exist anywhere else in the states, with such a concentration; we need to capitalize on it.”
Yet there are challenges, too. Many entrepreneurs, researchers and medical professionals are having to wrestle with the potential risks that come with the benefits of innovation. Some are proceeding cautiously as they navigate uncharted territories.
Chief among those concerns are data privacy, biases in algorithms, and the potential repercussions and ethics of handing over some medical decisions to technology that could have undetected faults.
For those reasons, Care New England Health System’s journey into artificial intelligence has been measured, according to Tomas Gregorio, CNE’s chief information officer.
So far, the health system has deployed AI to assist with administrative duties. Using it in medical settings is still far away, Gregorio says. More research needs to be done before hospital officials have that much confidence in the technology.
“We’re taking small bites of the apple,” he said.
Meanwhile, New Bedford-based hospital group Southcoast Health System Inc. – which provides some services in Rhode Island – has been nationally recognized for its use of AI in assisting with diagnostics and analyzing images, and Brown University Health – Rhode Island’s largest hospital group – has also been heavily involved in AI research and development.
“AI is everywhere, and I would happily and substantially back up that the AI is probably most valuable within a health care setting,” Petrie said.
[caption id="attachment_485375" align="aligncenter" width="1024"]

ON A ROLL: Severence MacLaughlin, founder of DeLorean Artificial Intelligence, at a presentation in Las Vegas in September 2024. The predictive health care technology company has numerous clients nationwide, including UnitedHealthcare Inc.
COURTESY DELOREAN ARTIFICIAL INTELLIGENCE[/caption]
EASING FRUSTRATIONS
At least initially, the highest value of AI may lie in saving doctors from burnout, a growing concern in the health care sector as physicians are required to collect and evaluate more patient data than ever for electronic health records.
“There’s just so much data and not enough time,” Petrie said.
That’s where AI stands out, with its ability to organize, analyze, sort and filter information nearly instantaneously, and potentially make predictions about future health outcomes, too.
MacLaughlin says that’s what DeLorean AI does.
The technology “trains” on patient data collected from electronic medical records, as well as health care system and hospitalization records, then uses that knowledge to provide doctors with predictions on whether certain patients will have a medical issue such as a heart attack.
It can also tell doctors if a patient’s condition is expected to worsen or if they are likely to get a disease, MacLaughlin says. Then the technology provides the doctors with recommendations for the best actions to help their patients.
According to MacLaughlin, testing has shown Delorean AI has been able to predict if a patient is going to have a heart attack within six weeks with 88% accuracy. And this will only improve as the company analyzes more patients’ data and improves its models, MacLaughlin says.
But at this point, AI is more often used in helping doctors with mundane but necessary jobs.
At Care New England, physicians have been using an “ambient intelligence” tool known as DAX, or Dragon Ambient eXperience, Gregorio says. This technology can record visits and automatically transcribe conversations between patients and doctors – cutting down on the time doctors spend taking notes.
The technology is offered through Microsoft Corp., which has collaborated with Epic Systems Corp. – the electronic medical records system to which Care New England is transitioning – to integrate large language modeling and other AI. Gregorio says all CNE facilities are expected to be using Epic by October.
Two medical residents at Brown University Health found a much more elementary use for AI.
When Drs. Rohaid Ali and Fatima Mirza found patients were struggling to make sense of complicated consent forms they must sign before medical procedures, they turned to ChatGPT, a web-based AI language model that can communicate conversationally with humans.
The doctors learned the forms had been written at a college reading level, even though most Americans read at a level much lower.
They tasked ChatGPT to come up with a simpler form while keeping the message and information. Now the ChatGPT-generated consent forms have been used for more than 40,000 procedures performed over the past year at Brown Health facilities.
Then there are the back-office jobs.
A group of students at Brown University have launched a startup, Kyron Medical, that uses AI to streamline the medical billing process, a major source of frustration for physicians.
Jay Gopal, who founded Kyron with fellow student Lucas Lieberman, says the idea came after cold-calling 200 doctors across the country to ask them about their workflows and what could be improved.
Many lamented the long hours spent doing mundane billing tasks and specifically handling claim denials from health insurance companies.
Their software uses large language modeling to automatically complete billing forms in the system already being used by a doctor, then can also appeal insurance claims that have been denied.
Kyron is still in its early stages, but Gopal says several customers have signed up to use Kyron and they hope to finalize the product after getting feedback from early adopters. The aim is to give physicians more time with their patients in the examination room.
That’s where Dr. Gaurav Choudhary doesn’t see AI replacing doctors.
Choudhary, director of cardiovascular research at Brown’s Warren Alpert Medical School, says the technology will likely play an increasing role in the way doctors diagnose illnesses, and will assist humans in reaching conclusions.
“The human relationship between a physician and patient will always be the centerpiece of what we do,” Choudhary said.
His research is focused on using AI algorithms to analyze digitally recorded heart sounds for early and hard-to-detect signs of pulmonary hypertension – or high blood pressure in the lungs – which affects about 1% of the world’s population.
Choudhary is in the second year of a three-year study funded by the National Institutes of Health.
In the future, the hope is AI will be used to create more personalized treatment plans.
Mirza and Ali got an early start on this by using Open AI’s Voice Engine to help a patient get her voice back last year.
The patient, Lexi Bogan, 21, had lost the ability to speak after doctors removed a brain tumor. But they were able to re-create Lexi’s voice in March by using a 15-second clip of her talking when she was a high school teen. Once that clip was fed into Voice Engine, it cloned her voice.
They were then able to give Bogan a custom-built iPhone app on which she types sentences that are read by the app in a synthetic voice that sounds just like hers once did.
“The best part was when she said ‘I feel like I have my voice back,’ ” Mirza said. “At the heart of this work with AI is empowering patients.”
NOT PERFECT
Gregorio acknowledges Care New England has been slower than other health care organizations to move toward AI. This is mainly because larger health systems have teams focused on researching how best to use AI outside the office, while Care New England does not.
At the same time, however, the state’s second-largest hospital group has been focusing on its transition to the Epic records system, which has AI features.
Care New England has formed a 12-member committee that will meet monthly with the task of examining how AI has been used in health care diagnostics and procedures. But until there’s more research into using AI in diagnostics and medical procedures, the health system will not proceed.
“Once those things become standard of practice, then we’ll see,” Gregorio said. “It’s too much too soon for us.”
He noted that the health system wants to ensure there are no biases built into any models it uses.
Studies have found that AI systems can suffer from bias – or results that are skewed because the information isn’t representative of all patients – adding to inequities in factors such as socioeconomic status, race, gender and sexual orientation.
The best way to prevent biases in AI models is to gather more data on which the algorithms can train, Choudhary says. But this takes time.
In his case, it will still take years to develop a reliable artificial intelligence for early detection of pulmonary hypertension. Data collection is expected to end next year. After that, it will take about six months for it to be analyzed. Then work on the AI algorithm can begin and it may be ready in a few years.
As powerful as AI is, it’s meant to be a tool that assists health care professionals – not fully replace them – and the technology requires consistent monitoring, Mirza says.
“We need to make sure the [AI] tools are supervised, and there are lots of eyes on it,” Mirza said.
But gathering large amounts of patient data raises questions about security, and unauthorized access to health information that is individually identifiable.
“Patient privacy is going to become a really big deal,” said James R. Langevin, the former congressman who is now the distinguished chair at Rhode Island College’s Institute for Cybersecurity & Emerging Technologies.
Privacy is clearly top of mind when the Brown students running Kyron Medical make their pitch to prospects. Questions usually follow about the protection of patient information and compliance with the Health Insurance Portability and Accountability Act, commonly known as HIPAA.
Gopal’s response: The Kryon software meets HIPAA requirements and patient data is encrypted from prying eyes.
The answer is similar for MacLaughlin at DeLorean AI when prospects ask about security. All 1 billion pieces of patient data used to train his predictive technology are safely behind a firewall, he says.
“From the beginning, privacy was a cornerstone,” MacLaughlin said.
[caption id="attachment_485374" align="alignright" width="416"]

PROBLEM SOLVING: Joshua Kemp, a clinical psychologist, has created a digital platform called CareCrowd Inc. that uses artificial intelligence to help users match with appropriate health providers.
COURTESY CARECROWD INC.[/caption]
FERTILE GROUND
There’s no shortage of Rhode Island entrepreneurs eager to get their AI technologies safely to market. And those in the sector expect more innovations to keep flowing from local higher education institutions, and from elsewhere.
Joshua Kemp is giving it a go.
Kemp, a clinical psychologist, launched CareCrowd Inc. months ago to help patients search for a health care provider.
Kemp says he and his colleagues often heard from patients who spent months struggling to find a provider because it’s not clear where to look, who is available and whether that provider would be a good fit.
Also, care coordinators responsible for assisting patients with their health care needs told Kemp they spend up to 40% of their time searching through provider directories that are often filled with wrong information.
“It’s an open secret that some of the data out there is not very accurate,” Kemp said.
CareCrowd uses AI algorithms to train on publicly available data, combined with crowd-sourcing information on available providers, to create a directory of providers that is much easier for patients to navigate.
Kemp launched an early version of the product with a small number of care coordinators who signed up to test it. He’ll use that feedback to create a fine-tuned version.
But Kemp says he’s struggled to find investors willing to fund the product. So far there’s been interest in seed funding, but investors wanted to see a version of the product in action before signing any checks.
Kemp isn’t alone at this stage.
Petrie says funding is one of several roadblocks faced by many creating AI products for the health care market.
Especially if the product needs to get approved by the U.S. Food and Drug Administration. In that case, there needs to be a clinical trial, which could take months and costs hundreds of thousands of dollars, Petrie says. There’s also HIPAA compliance.
“The path is a well-worn road, and we just have to have a good plan, and you’ve got to convince people to give you lots of money to then bring it to market,” Petrie said.
In the meantime, Petrie says, Rhode Island is positioned to assist entrepreneurs because its smallness can lead to collaboration and access to key players.
“I can walk out of my office and I have three engineering schools and a medical school within 200 yards,” said Petrie, whose office is located in Providence’s Jewelry District.
MacLaughlin says networking and persistence helped get Delorean AI into 430 renal clinics across the U.S. – including several in Rhode Island – and there are now plans for expanding to 230 nephrology clinics across the U.S.
Along with that, MacLaughlin said several countries have expressed interest in Delorean and clients such as UnitedHealthcare have requested new AI predictive modules for high-risk pregnancies and musculoskeletal diseases.
“I just decided to put the pedal to the metal,” MacLaughlin said. “The goal is to help people.”