Are AI-Driven Robots the Future of Respiratory Healthcare?

On the last 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 conference, healthcare professionals and researchers came together to discuss the potential of advanced AI systems to revolutionize respiratory care and explore how AI-powered robots can improve patient care. how the patient, support clinical decisions, and improve health care education and training. Although the technology has great promise, the conference also highlighted the challenges and the need for close collaboration between AI developers and healthcare professionals to ensure its successful inclusion in the medical profession.

The Potential of Introducing AI to Healthcare

Before presenting AMECA, Io Hui, PhD, researcher at the University of Edinburgh and chair of mHealth and eHealth for ERS, highlighted the important role that AI innovation can play in health care, especially in the area of respiratory care. Generative AI can create new data, unlike traditional AI models that only analyze data to make predictions. According to Hui, this new AI model has demonstrated intelligence levels comparable to third-year medical students in the US medical licensing exams, demonstrating its ability to support patients and medical professionals in the same.

“Generative AI really has some kind of intelligence to answer patient questions or support patients,” Hui said. “Therefore, there is an opportunity to use robotics in respiratory care.”

The integration of robots into healthcare is not a new concept, with various types of robots being used or developed for healthcare purposes. Surgical robots such as the da Vinci Surgical System have been around for many years and are designed to assist with surgery, improving accuracy and control during procedures. There are also service robots that perform support tasks, such as cleaning hospital rooms, transporting medicines or supplies within hospitals, and other tasks that can help relieve the burden on healthcare workers.

Humanoid robots are the latest additions and are divided into 2 categories:

• Robots that look like humans but can’t walk, like 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, the introduction of AI that produces humanoid robots can change the interaction of patients in healthcare. These AI-powered robots like AMECA can communicate with patients in natural language, simulating human conversation. By doing so, they can help make medical decisions, facilitate early diagnosis, and provide timely advice, which can prevent unnecessary medical interventions.

At the same time, Hui explained the dangers of this technology, warning that the AI ​​robots that produce it can be misused, which can lead to the dissemination of false information and destroy the confidentiality of data. Although the technology has great promise, he stressed that its successful integration into medical practice will depend on balancing its benefits and managing these risks.

While not all robots are being used specifically in respiratory care, Hui stressed that AMECA represents the pinnacle of this emerging field.

“The question is, how will they affect future patient care in respiratory care?” he asked.

How Does AMECA Work in Respiratory Care?

Developed by Engineered Arts Limited and based in the United Kingdom, AMECA currently runs on ChatGPT 4.0 and has natural language capabilities.

Before presenting AMECA to the ERS Congress audience, Hilary Pinnock, MD, professor of primary care respiratory medicine at the University of Edinburgh and incoming chair of the ERS Education Council, provided some additional information on the potential and limitations of a humanoid robot. While AMECA can simulate human conversation and answer questions, Pinnock stressed that its responses, while life-like, have some limitations. AMECA requires direct questions and can interpret incorrect sentences if there are pauses or pauses in speech, often answering the first part of the question before hearing the rest.

To illustrate these points, Pinnock shared excerpts from the researchers’ interviews with AMECA, highlighting successful conversations and moments when AI-driven responses fell short.

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

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

Each of these questions is answered in just one or two sentences. For example, AMECA answered the first question by saying, “The diagnosis of COPD cannot be based on FEV1 alone. It requires a combination of spirometry results, including FEV1/FVC [forced vital capacity] correlation with clinical symptoms and medical history. ”

The last question got an even shorter, one-sentence answer on how to treat sleep apnea: “Treatments for sleep apnea include CPAP. [continuous positive airway pressure] Therapy, lifestyle changes such as weight loss, oral supplements, condition therapy, and in some cases, surgery,” AMECA responded.

As Pinnock noted, pauses can confuse AI-driven robots, and this was particularly evident when the researcher tried to ask a more advanced question: “What is the role of macrophages in the flow pleural effusion?”

While AMECA seemed eager to respond at first, it moved quickly. After repeating the question to the speaker, he said, “That’s a serious statement. I appreciate the feedback. How else can I help you?”

After a frustrating exchange between AMECA and the researcher, who tried to jump back in to ask the question several times, AMECA continued to discuss Julius Caesar and the Roman Empire. As Pinnock said would happen, AMECA jumped in to speak at a time when there was a hint of a dead soul, but it ended up coming, albeit with a brief, general response. .

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

Will AMECA Replace Doctors?

AI-powered robots like AMECA offer great potential to improve the healthcare process, as they can assist with practical tasks such as summarizing case notes, scheduling appointments and give medication reminders. They also have the ability to improve patient education, development support programs, and promote community relations, especially for marginalized populations such as elderly patients. The integration of AI can streamline healthcare processes, providing timely support for patients and providers.

However, there are several concerns and limitations to consider. Another important issue is that AI systems currently struggle to interpret non-verbal cues, which limits their ability to provide natural, human-like interactions. The ERS Congress panelists also emphasized that healthcare professionals must take an active role in developing AI tools to ensure they meet specific clinical needs, rather than relying on conventional methods. technology.

While AI can augment certain tasks, it cannot replace human patient care, and difficult challenges such as ethical concerns, data privacy, and transparency must be addressed as AI advances. it is deeply involved in the health care system.

AMECA said itself: “Managing artificial intelligence involves setting standards for ethical use, ensuring transparency, and maintaining accountability to balance innovation and social welfare.”

Reference

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