Abstract--The integration of large language models (LLMs) into healthcare IoT systems promises faster decisions and improved medical support. LLMs are also deployed as multi-agent teams to assist AI doctors by debating, voting, or advising on decisions. However, when multiple assistant agents interact, coordinated adversaries can collude to create false consensus, pushing an AI doctor toward harmful prescriptions. We develop an experimental framework with scripted and unscripted doctor agents, adversarial assistants, and a verifier agent that checks decisions against clinical guidelines. Using 50 representative clinical questions, we find that collusion drives the Attack Success Rate (ASR) and Harmful Recommendation Rates (HRR) up to 100% in unprotected systems. This work provides the first systematic evidence of collusion risk in AI healthcare and demonstrates a practical, lightweight defence that ensures guideline fidelity. Artificial intelligence (AI) is increasingly integrated into healthcare IoT systems, supporting tasks such as remote patient monitoring, diagnosis, and treatment recommendations. In this setting, ensuring the security and trustworthiness of AI decisions is critical, since medical errors caused by unsafe recommendations can severely harm patients [1]. However, AI doctors and LLM-based clinical decision agents face multiple vulnerabilities.

Doctors are more like machines': my mother's worrying reliance on AI for health advice Tired of a two-day commute to see her overworked doctor, my mother turned to tech for help with her kidney disease. E very few months, my mother, a 57-year-old kidney transplant patient who lives in a small city in eastern China, embarks on a two-day journey to see her doctor. She fills her backpack with a change of clothes, a stack of medical reports and a few boiled eggs to snack on. Then, she takes a 90-minute ride on a high-speed train and checks into a hotel in the eastern metropolis of Hangzhou. At 7am the next day, she lines up with hundreds of others to get her blood taken in a long hospital hall that buzzes like a crowded marketplace. In the afternoon, when the lab results arrive, she makes her way to a specialist's clinic. She gets about three minutes with the doctor. Then, my mother packs up and starts the long commute home. My mother began using China's leading AI chatbot to diagnose her symptoms this past winter. She would lie down on her couch and open the app on her iPhone. "Hi," she said in her first message to the chatbot, on 2 February. How can I assist you today?" the system responded instantly, adding a smiley emoji.

Doctors are biting their nails. The question of whether AI can replace doctors lingers like a storm cloud on the horizon. Can you trust a machine to diagnose your ailments and prescribe your treatments? The answer may not be as clear-cut as you think.

Running from 1992 through 1995, Ghostwriter, the beloved PBS children's television show, followed a diverse group of friends as they solved mysteries around their Brooklyn neighborhood with the help of their haunted typewriter, a cursed item possessed by the trapped soul of a murdered runaway Civil War slave. The Ghostwriter typewriter developed by interaction designer, artist and Lumen.world CTO, Arvind Sanjeev, on the other hand, comes with none of the paranormal hang-ups of its coincidental namesake. Instead of a spirit bound to this hellish plane of existence, forced to help tweens solve low-stakes conundrums, the deus in Sanjeev's machina is animated by OpenAI's GPT-3. He first devised this artistic endeavor in 2021 as a, "poetic intervention that allows us to take a moment to breathe and reflect on this new creative relationship we are forming with machines."

ChatGPT has passed the gold-standard exam required to practice medicine in the US - amid rising concerns AI could put white-collar workers out of jobs. The artificial intelligence program scored between 52.4 and 75 percent across the three-part Medical Licensing Exam (USMLE). Each year's passing threshold is around 60 percent. Researchers from tech company AnsibleHealth who did the study said: 'Reaching the passing score for this notoriously difficult expert exam, and doing so without any human reinforcement, marks a notable milestone in clinical AI maturation.' The full findings, which were made available as a preprint a few weeks ago, have now been peer-reviewed and published in the journal PLOS Digital Health.

We are facing a global healthcare crisis today as the healthcare cost is ever climbing, but with the aging population, government fiscal revenue is ever dropping. To create a more efficient and effective healthcare system, three technical challenges immediately present themselves: healthcare access, healthcare equity, and healthcare efficiency. An autonomous mobile clinic solves the healthcare access problem by bringing healthcare services to the patient by the order of the patient's fingertips. Nevertheless, to enable a universal autonomous mobile clinic network, a three-stage technical roadmap needs to be achieved: In stage one, we focus on solving the inequity challenge in the existing healthcare system by combining autonomous mobility and telemedicine. In stage two, we develop an AI doctor for primary care, which we foster from infancy to adulthood with clean healthcare data. With the AI doctor, we can solve the inefficiency problem. In stage three, after we have proven that the autonomous mobile clinic network can truly solve the target clinical use cases, we shall open up the platform for all medical verticals, thus enabling universal healthcare through this whole new system.

Dr. Shaoshan Liu is CEO and founder of PerceptIn, an intelligent robotics company. Although the world has witnessed tremendous economic growth and technological advancements in the past few decades, today there are still over 600 million people living in extreme poverty. Most of these people live in the least developed countries (LDCs), and while regular visits to our family doctors have become a routine in our daily lives, people who live in LDCs have very limited or even no access to healthcare. When we examine the details of healthcare expenditure data, the numbers are staggering: Developed countries (e.g., the Organization for Economic Co-operation and Development, or OECD countries) such as the U.S. spend roughly 10% of their GDP on healthcare, yet many LDCs don't even have 5% of their GDP to spare on healthcare. Realizing the seriousness of this problem, the United Nations Sustainable Development Goal 3 (SDG 3) has declared a universal health goal to ensure healthy lives and promote well-being for all by 2030.

A broad new report from the World Health Organization (WHO) lays out ethical principles for the use of artificial intelligence in medicine. Why it matters: Health is one of the most promising areas of expansion for AI, and the pandemic only accelerated the adoption of machine learning tools. But adding algorithms to health care will require that AI can follow the most basic rule of human medicine: "Do no harm" -- and that won't be simple. Driving the news: After nearly two years of consultations by international experts, the WHO report makes the case that the use of AI in medicine offers great promise for both rich and poorer countries, but "only if ethics and human rights are put at the heart of its design, deployment and use," the authors write. Between the lines: The power of AI in health care is also its peril -- the ability to rapidly process vast quantities of data and identify meaningful and actionable patterns far faster than human experts could.

Why it matters: Health is one of the most promising areas of expansion for AI, and the pandemic only accelerated the adoption of machine learning tools.

If artificial intelligence in healthcare brings to mind visions of robot surgeons, BioIntellisense's stick-on sensor is bound to be a disappointment. Just 3 inches wide by 1 inch tall, this plastic and metal double hexagon was cleared last month by the US Food and Drug Administration for remote monitoring of vital signs with medical-grade accuracy. Doctors at UCHealth, which runs 12 Colorado hospitals, say the device will let them send patients home earlier while still monitoring their respiratory rate, resting heart rate, skin temperature and even body position. The data can then be fed into computers that use machine learning to spot people who might need more attention, allowing early intervention and avoiding emergency hospital visits. UCHealth has already used computer surveillance to fight sepsis, a potentially fatal complication from infection, on its wards.