The Accelerating State Of AI Health In Hospitals And Homes

Forbes

27-04-2025

18 July 2024, Brandenburg, Cottbus: A live interaction between a simulated patient (Doris Härtel) ... More and a robot can be seen at a press event at the Carl-Thiem-Klinikum Cottbus. A pilot study is currently being carried out at the Brandenburg University of Technology Cottbus-Senftenberg (BTU) to investigate the effects of the interaction modalities of a social robot (Pepper) on the engagement and perception of test subjects in typical care situations. Photo: Patrick Pleul/dpa (Photo by Patrick Pleul/picture alliance via Getty Images)
The adoption of AI tools to improve the quality and cost-effectiveness of healthcare is accelerating. The impact is felt in multiple domains: from developing a drug or treatment to deploying it, assisting with managing patients' health in clinical settings, and putting AI in the hands of consumers. Here is a roundup of recent growth indicators of AI in healthcare.
In the first quarter of 2025, we saw a marked acceleration in AI health investments. CB Insights reports that AI startups raised $3.2 billion, or 60% of all digital health funding, up from 41% in the first quarter of 2024. 'Top-funded segments included AI-derived small molecule drug discovery and clinical documentation tools, underscoring the shift toward targeted, high-impact applications,' says CBI. AI startups secured 8 of the 11 mega-rounds (deals over $100 million), 'signaling where investors expect outsized returns.'
A recent study by Yijin Hardware found 11,228 healthcare-specific active AI startups worldwide. Based on the extensive investment in these startups, the number of industry-specific AI startups, interest in AI technologies reflected through Google search volume, and top uses for AI technologies in the industry, the study declared healthcare as 'the most AI-driven industry in 2025.'
I wrote before about the significant segment of the more than 11,000 AI health startups, those that are involved with drug and treatment development. For example, Nucleai, working with pharmaceutical companies to improve the process of oncology drug development, and OncoHost, a developer of a proteomic analysis tool that guides decision-making in the choice of first-line treatment for cancer patients. Ofer Sharon, CEO of OncoHost, wrote recently that 'in 2025, we stand on the cusp of a new era where AI technologies… are not only enhancing drug discovery and diagnostics but also driving broader innovations that improve patient outcomes via personalized approach to complex disease management.'
Here, I focus on recent examples of AI's impact on healthcare providers' work and on consumers' management of their health and well-being.
According to CBI, investors are particularly excited about AI managing the workflow in hospitals and clinics. Half of the 6 new digital health unicorns minted in the first quarter of 2025 (more unicorns than in all of last year) are focused on this area.
This is a crucial area for improving physicians' and patients' well-being. I will write more in the future about how AI is easing the administrative burden associated with navigating the complex rules and cumbersome processes typical of the healthcare sector. Here, I focus on a handful of indicators of AI's current and potential involvement with efficient healthcare management and patient diagnosis and treatment decisions.
At Sheba Medical Center in Israel, dozens of patients in a pilot in the emergency department interact with an AI medical agent. The agent automatically compiles comprehensive health summaries for doctors, recommends imaging and laboratory tests, and provides clinical decision support—saving time, reducing paperwork, improving diagnostic accuracy, and enhancing the patient experience. This allows doctors to manage multiple patients simultaneously and prioritize those in critical condition.
Viz.ai's AI-based stroke detection system analyzes CT scans in real time to identify critical conditions such as strokes, aneurysms, and pulmonary embolisms. It is used in more than 1,700 hospitals in the U.S. and Europe. In Q1 2025, Viz.ai released a new version, demonstrating improved accuracy and faster detection times. This system saves lives by enabling hospitals to initiate treatment protocols more rapidly.
India's Apollo Hospitals, one of the largest hospital networks in the country, set aside 3.5% of its digital spend on AI over the past two years. This year, it plans to increase it 'to free up two to three hours of time daily for doctors and nurses with AI interventions." These include analysis of patients' electronic medical records to suggest diagnoses, tests, and treatment. In addition, AI helps transcribe doctors' observations, generate faster discharge summaries, and create daily schedules out of nurses' notes.
Keiju General Hospital in Japan uses AI discharge summary tools developed by Ubie, a local startup. These tools reduce nurses' time on these tasks by 42.5% and decrease their psychological burden by 27.2%. In a recent trial at Kyushu University Hospital, one of the largest hospitals in Japan, summarizing and standardizing referral letters led to a 54% increase in efficiency for doctors preparing admission summaries.
Mercy Health's AI program reduced nurse charting time by 34 minutes per shift, from 167 to 133 minutes, addressing a key pain point. The health system's ambient AI pilot, now nearly three years old, has proven critical in streamlining workflows while maintaining care quality.
At the ViVE Digital Health conference in February, Dr. John Halamka, president of the Mayo Clinic Platform, highlighted the growing use of ambient listening tools. Health systems are using these tools to document patient visits, allowing doctors to have natural conversations with patients rather than typing notes on a computer. These AI tools also provide summaries of the patient encounter, enabling clinicians to save time and energy. Mayo Clinic uses an inpatient ambient nursing solution in Arizona and Florida that does '100% of the nursing charting without the nurse having to touch a keyboard.'
AI is also becoming involved in assisting physicians with complex treatment decisions. For example, the Princess Máxima Center for Pediatric Oncology in the Netherlands is developing an AI tool to help physicians accelerate the identification of personalized cancer treatments by combining vast public medical data and de-identified patient data. Based on its analysis, the tool rapidly generates summaries of treatment options and the relevant medical sources.
The National Institutes of Health released on April 3 a study that found that an artificial intelligence screening tool was as effective as health care providers in identifying hospitalized adults at risk for opioid use disorder and referring them to inpatient addiction specialists. Compared to patients who received consultations with providers, patients screened by AI had 47% lower odds of hospital readmission within 30 days after their initial discharge, saving nearly $109,000 in care costs.
Over the last decade, numerous studies have compared the accuracy of AI-based medical diagnosis to that of medical experts in different specialties. Increasingly, we learn about hospitals moving to deploy such AI-based diagnostic models.
University Hospitals Cleveland Medical Center announced in early April a collaboration with Qure.ai to deploy the FDA-cleared chest X-ray AI solution qXR-LN, which will support earlier identification of lung cancers. The solution will act as a second read, to be compared to the radiologists' read of patient chest X-rays for any suspicious lung nodules.
Lunit's advanced mammography system, which leverages AI to detect breast cancer, can now detect breast cancer up to six years earlier than traditional methods, significantly improving early intervention and treatment outcomes. Google's DeepMind division recently released an updated version of their AI model, which has shown improved accuracy in detecting subtle signs of breast cancer that human radiologists might miss.
According to a new study by startup K Health and researchers at Cedars-Sinai and Tel Aviv University, AI's medical decisions can sometimes be better than those of human doctors. The study found that K Health's AI chatbot, which makes recommendations and diagnoses based on the patient's medical records and conversations with the patient, matched the doctor's decisions in two-thirds of cases and offered better-quality care in the remaining one-third.
58% of Americans used virtual care in the past year. Healthcare organizations that implement AI-enabled telehealth report substantial improvements across multiple metrics. About 75% of facilities note enhanced disease treatment effectiveness, while 80% experience reduced staff burnout rates. AI systems process medical images 30 times faster than conventional methods without sacrificing accuracy. Cleveland Clinic reports a 94% accuracy rate for its AI virtual triage system.
53% of consumers own at least one wearable, and 54% track at least one health metric digitally. In February, Google Health received clearance from the FDA for its Loss of Pulse Detection feature on the Pixel Watch 3. This first-of-its-kind feature can detect a loss of pulse (when the heart stops beating from an event like primary cardiac arrest, respiratory or circulatory failure, overdose, or poisoning) and automatically call emergency services.
And more to come… Samsung and Stanford Medicine announced a joint research initiative to enhance the Galaxy Watch's sleep apnea detection feature and create new AI-enabled innovations for proactive care. Apple announced its planned Health+ app, which includes an AI-enabled health coach that shares personalized health insights derived from wearable device data.
In addition to investors, startups, and healthcare providers, the acceleration of AI health adoption has reached the federal government. As of October 2024, the FDA approved 1,000 AI-enabled medical devices. It approved just six in 2015, 160 in 2022, and 223 in 2023.