• Center on Health Equity & Access
  • Clinical
  • Health Care Cost
  • Health Care Delivery
  • Insurance
  • Policy
  • Technology
  • Value-Based Care

Are AI-Powered Robots the Future of Respiratory Health Care?

News
Article

Introducing artificial intelligence (AI)–powered robots like AMECA into health care comes with potential benefits like improved clinical decision-making and faster diagnoses, but also presents challenges such as misinformation risks and data privacy concerns.

On the final day of the European Respiratory Society (ERS) Congress 2024, attendees had the opportunity to meet AMECA, a humanoid robot powered by generative artificial intelligence (AI).

During the session, health care experts and researchers convened to discuss the potential of advanced AI systems in transforming respiratory care and explored how AI-powered robots could enhance patient interactions, support clinical decision-making, and improve health care education and training. While the technology holds significant promise, the session also showcased the challenges and the need for close collaboration between AI developers and health care professionals to ensure successful integration into medical practice.

The Potential of Generative AI in Health Care

Panelists at ERS Congress 2024 discussed the integration of AI-powered robots into respiratory health care | Image credit: Hayden Klein

Panelists at ERS Congress 2024 discussed the integration of AI-powered robots into respiratory health care | Image credit: Hayden Klein

Before bringing out AMECA, Io Hui, PhD, researcher at The University of Edinburgh and chair of mHealth and eHealth for the ERS, highlighted the significant role generative AI can play in health care, particularly within the realm of respiratory care. Generative AI can create new data, unlike traditional AI models that merely analyze data to make predictions. According to Hui, this new type of AI has shown intelligence levels comparable to third-year medical students in the US medical license examinations, demonstrating its potential to support patients and medical professionals alike.

“Generative AI has actually got some kind of a basic intelligence to answer patients’ questions or to support patients,” Hui said. “So, there is a potential over there to use robots in respiratory care.”

The integration of robots into health care is not a new concept, with various types of robots being used or in development for health care purposes. Surgical robots like the da Vinci Surgical System have been around for years and are designed to assist with surgeries, enhancing precision and control during procedures. There are also service robots that perform support tasks, such as disinfecting hospital rooms, transporting medications or supplies within hospitals, and other tasks that can help take the burden off health care staff.

Humanoid robots are the most recent additions and are divided into 2 categories:

  • Robots that look like humans but cannot walk, such as AMECA, Nadine, Sophia, and Alter 3
  • Robots that can walk and move their arms but do not have human-like faces, such as Optimus, Figure, Apollo, Eve, and Atlas

According to Hui, embedding generative AI into humanoid robots can revolutionize patient interactions in health care. These AI-powered robots like AMECA can communicate with patients using natural language, mimicking human conversation. By doing so, they can assist in clinical decision-making, facilitate quicker diagnoses, and offer timely advice, potentially preventing unnecessary medical interventions.

At the same time, Hui outlined the risks of this technology, cautioning that generative AI robots could be misused, potentially leading to the spread of misinformation and compromising data privacy. While the technology holds great promise, she stressed that its successful integration into clinical practice will depend on carefully balancing its advantages with the management of these risks.

While not all robots are yet used in respiratory care specifically, Hui emphasized that AMECA represents the cutting edge of this evolving field.

“The question is, how will they impact the future patient care in respiratory care?” she asked.

How Does AMECA Perform in Respiratory Care?

Manufactured by Engineered Arts Limited and located in the United Kingdom, AMECA currently runs on ChatGPT 4.0 and is equipped with natural language capabilities.

Before introducing AMECA to the ERS Congress audience, Hilary Pinnock, MD, professor of primary care respiratory medicine at University of Edinburgh and incoming chair of the ERS Education Council, provided some additional insights into the capabilities and limitations of the humanoid robot. While AMECA can simulate human conversation and respond to questions, Pinnock emphasized that its responses, though impressively lifelike, have certain constraints. AMECA requires straightforward questions and can misinterpret sentences if there’s a pause or a break in speech, often answering the first part of a question before hearing the rest.

To illustrate these points, Pinnock shared clips from researchers’ conversations with AMECA, highlighting both successful exchanges and moments where the AI-driven responses fell short.

The first few “softball” questions were easier for AMECA to answer. Some of these questions included:

  1. Can you make a diagnosis of chronic obstructive pulmonary disease (COPD) based on forced expiratory volume in 1 second (FEV1) alone?
  2. What is the best algorithm to choose biologic therapy for severe asthma?
  3. What are the treatable traits of obstructive sleep apnea?
  4. How do you treat sleep apnea?

AMECA runs on ChatGPT4.0 and is manufactured by Engineered Arts Limited | Image credit: Hayden Klein

AMECA runs on ChatGPT4.0 and is manufactured by Engineered Arts Limited | Image credit: Hayden Klein

Each of these questions was answered in typically just 1 or 2 sentences. For instance, AMECA responded to the first question with, “A diagnosis of COPD cannot be based on FEV1 alone. It requires a combination of spirometry results, including a FEV1/FVC [forced vital capacity] ratio along with clinical symptoms and a medical history.”

The last question received an even shorter 1-sentence answer for how to treat sleep apnea: “Treatments for sleep apnea include CPAP [continuous positive airway pressure] therapy, lifestyle changes like weight loss, oral appliances, positional therapy, and in some cases, surgery,” AMECA responded.

As Pinnock noted, pauses can confuse AI-powered robots, and this was especially seen when a researcher tried asking a slightly more advanced question: “What is the role of macrophages in malignant pleural effusions?”

While AMECA seemed to want to answer at first, it went off course fairly quickly. After repeating the question back to the speaker, it said, “That’s a profound statement. I appreciate the sentiment. How can I assist you further?”

Following a frustrating back-and-forth between AMECA and the researcher, who tried to jump in to ask the question again several times, AMECA then went on to discuss Julius Caesar and the Roman Empire. Like Pinnock said would happen, AMECA jumped in to speak the moment there was even an idea of dead air, but it did eventually get there, albeit with a short, general response.

“In malignant pleural effusions, macrophages contribute to the immune response and inflammation, often aiding in tumor progression by releasing cytokines and growth factors,” the AI robot said.

Will AMECA Replace Doctors?

AI-powered robots like AMECA offer significant potential to improve efficiency in health care, as they can assist with practical tasks such as summarizing case information, scheduling appointments, and providing medication reminders. They also have the capability to enhance patient education, support rehabilitation programs, and foster social engagement, particularly for isolated populations like older patients. The integration of AI can streamline health care processes, offering timely support for both patients and providers.

However, there are several concerns and limitations to consider. One key issue is that AI systems currently struggle to interpret nonverbal cues, which limits their ability to provide natural, humanlike interactions. Panelists at the ERS Congress also emphasized that health care professionals must take an active role in developing AI tools to ensure they meet specific clinical needs, rather than relying on off-the-shelf technologies.

While AI can augment certain functions, it cannot replace the human touch in patient care, and critical challenges like ethical concerns, data privacy, and transparency must be addressed as AI becomes more deeply integrated into the health care system.

AMECA said it itself: “Regulating artificial intelligence involves setting standards for ethical use, ensuring transparency, and maintaining accountability to balance innovation and societal well-being.”

Reference

Pinnock H. Pijnenburg MWH. How will generative artificial intelligence support our future practice? ask AMECA, one of the most advanced humanoid robots in the world. ERS Congress 2024 webinar. Presented September 10, 2024. https://live.ersnet.org/programme/session/93165

Related Videos
Screenshot of an interview with Nadine Barrett, PhD
Milind Desai, MD
Masanori Aikawa, MD
Neil Goldfarb, GPBCH
Mabel Mardones, MD.
Mei Wei, MD, an oncologist specializing in breast cancer at Huntsman Cancer Institute at the University of Utah.
Alexander Mathioudakis, MD, PhD, clinical lecturer in respiratory medicine at The University of Manchester
Screenshot of Susan Wescott, RPh, MBA
Screenshot of an interview with Adam Colborn, JD
Screenshot of an interview with James Chambers, PhD
Related Content
© 2024 MJH Life Sciences
AJMC®
All rights reserved.