Change the conversation in healthcare | MIT News

Generative AI is changing the way humans write, read, think, empathize and act in language and culture, and across languages ​​and cultures. In health care, the communication gap between patients and practitioners can worsen patient outcomes and prevent improvements in practice and care. Language/AI incubators are achieved through funding from MIT Human Insight Collaboration (MITHIC), providing a potential response to these challenges.

The project envisions a research community rooted in the humanities that will foster interdisciplinary collaboration between MIT to deepen understanding of the impact of generative AI on cross-linguistic and cross-cultural communication. The project’s focus on healthcare and communication is aimed at building bridges across socio-economic, cultural and linguistic classes.

The incubator is co-led by Leo Celi, Physician and Research Director and Physician of the Institute of Medical Engineering and Science (IMES), Research Director and Senior Research Scientist, while Per urlaub, Professor of German and Second Language Research Practice and Director of the MIT Global Language Program.

“The foundation of providing health care is the understanding of health and disease,” Celi said. “We see bad results despite the rupture of our knowledge system, despite the substantial investment.”

Occasional cooperation

Urlaub and Celi meet in a miracle launch event. The conversations during the event reception show that people share a common interest in medical communication with AI and improvement in practice.

“We are trying to incorporate data science into healthcare,” Celi said. “We have been recruiting social scientists (at IMES) to help advance our work because the science we create is not neutral.”

The group believes that language is a non-neutral mediator for health care and can be a boon or barrier to effective treatment. “Later, after we met, I joined one of his working groups, whose working groups were the metaphor of pain: the language we used to describe it and the language we measured,” Urlaub continued. “One of the issues we consider is how effective communication between doctors and patients.”

They believe that technology affects casual communication, and its impact depends on users and creators. As AI and large language models (LLMS) gain power and prominence, their use is expanding to include areas such as health care and health.

Rodrigo Gameiro, a physician and researcher at MIT’s Computational Physiology Laboratory, is another program participant. He noted that he worked in the laboratory center responsible for AI development and implementation. Designing systems that effectively utilize AI, especially when considering challenges related to the possible linguistic and cultural divides in healthcare, requires a nuanced approach.

“When we build AI systems that interact with human language, we’re not just teaching machines how to process words; we’re teaching them to browse the complex network of meaning embedded in the language,” Gameiro said.

The complexity of language can affect treatment and patient care. Urlaub continues: “Pain can only be conveyed through metaphors, but metaphors do not always match linguistically and culturally.” Smiley faces and a pair of 10 scales – pain measurement tools that English medicine professionals may use to evaluate their patients – may not travel well on racial, racial, cultural and linguistic boundaries.

“Science must have a heart”

LLM can potentially help scientists improve healthcare, although some systematic and teaching challenges need to be considered. Celi believes that science can focus on excluding the outcomes of people it would have been about to help. “Science must have a heart,” he said. “The effectiveness of students is measured by counting the number of papers published or the number of patents produced.”

This is a careful investigation, Urlaub said, while recognizing what we don’t know, citing what philosophers call epistemological humility. Investigators believe that knowledge is temporary and always incomplete. Deep beliefs may need to be revised based on new evidence.

“No one’s psychological view of the world is complete,” Celi said. “You need to create an environment where people can comfortably acknowledge their biases.”

“How do we share our attention between language educators and others interested in AI?” Urlaub asked. “How do we identify and investigate the relationship between medical professionals and language educators who are interested in the potential of AI to help eliminate the communication gap between doctors and patients?”

According to Gameiro’s estimates, language is more than just a tool for communication. “It reflects culture, identity and power dynamics,” he said. Misunderstanding can be dangerous in situations where patients are reluctant to describe pain or discomfort because of their position as authority, or because of their cultural demands, succumb to the demands of people seen as authoritative figures.

Change the conversation

AI’s facilities that use language can help healthcare professionals browse these areas more carefully, thus providing digital frameworks that provide valuable cultural and linguistic environments where patients and practitioners can rely on data-driven, research-enabled research-enabled tools to improve conversations. The team said agencies need to rethink how they educate healthcare professionals and invite the communities they serve to the conversation.

“We need to ask ourselves. Why do we measure our own measurements?” Celi said. The bias we bring about by these interactions — doctors, patients, their families and their communities — remains a barrier to improving care, says ULab and Gameiro.

“We want to connect all kinds of thinking people together and make AI work for everyone,” Gameiro continues. “Purposeless technology is only excluded on a large scale.”

“This collaboration could allow for deep processing and better ideas,” Urlaub said.

Creating ideas about AI and healthcare can become a space for action is a key element of the project. The language/AI incubator held its first symposium at MIT in May, and the campus is led by Mena Ramos, physician and co-founder and CEO of the Global Institute of Ultrasound.

The symposium also provided Celi’s speech, as well as Alfred Spector, a visiting scholar in the Department of Electrical Engineering and Computer Science in MIT, and Douglas Jones, a senior staff member of the Human Language Technology Group at the Lincoln Laboratory. A second language/AI incubator symposium is planned in August.

A greater integration between social sciences and hard sciences may increase the likelihood of developing viable solutions and reducing bias. Allowing patients and doctors to change their perceptions of this relationship while providing each shared interaction ownership can help improve outcomes. Facilitating these conversations through AI may accelerate the integration of these perspectives.

“Community advocates have voices and should be included in these conversations,” Celi said. “AI and statistical modeling cannot collect the data needed by everyone who needs it.”

Community needs and improved educational opportunities and practices should be combined with interdisciplinary approaches to knowledge acquisition and transfer. The way people see things is limited by their perceptions and other factors. “Whose language are we modeling?” Gameiro asked to build LLMS. “What voices are included or excluded?” Because meanings and intentions can be transferred in these contexts, it is important to remember these situations when designing AI tools.

“AI is our opportunity to rewrite the rules”

Despite the many potential for collaboration, serious challenges to overcome include building and expanding technical means to improve patient communication with AI, expanding collaboration opportunities, thus expanding marginalized and underserved communities, and rethinking and rethinking and improving patient care.

But the team is not afraid.

Celi believes that while addressing the gap in health care, there is an opportunity to address the widening gap between people and practitioners. “Our purpose is to reconstitute the strings that cut between society and science,” he said. “We can enable scientists and the public to investigate the world together while also acknowledging the limitations of overcoming bias.”

Gameiro is a passionate advocate of AI’s ability to transform everything we can about medical knowledge. “I’m a doctor and I don’t think I’m exaggerating when I say I believe AI is an opportunity for us to rewrite the rules of what medicine can do and what we can achieve,” he said.

“Education turns humans from objects into objects,” Urlaub believes. The difference between selfless observers and active and involved participants in the new nursing model he hopes to build. “We need to better understand the impact of technology on the boundaries between these states of existence.”

Celi, Gameiro and Urlaub each advocate that trees throughout healthcare resemble wooden spaces that can innovate and collaborate without any arbitrary benchmarks that institutions have previously used to mark success.

“AI will change all these departments,” Urlaub believes. “Mithic is a generous framework that allows us to embrace uncertainty in a flexible way.”

“We want to use our power to build communities among different audiences while acknowledging that we don’t have all the answers,” Celi said. “If we fail, it’s because we don’t have big enough dreams to achieve the look of a reimagined world.”