AI as Normal Technology
How Generative AI Is Reshaping Intensive Care Medicine?
Advancements in artificial intelligence are reshaping the landscape of intensive care medicine, with generative AI models emerging as a critical technology for predicting patient outcomes. A team of researchers has provided the first comprehensive scoping review on this topic, evaluating the role of generative AI in critical care prognostication.
Published in Frontiers in Digital Health, the study titled "Applications of Generative Artificial Intelligence in Outcome Prediction in Intensive Care Medicine—A Scoping Review," the study analyzed hundreds of papers and identified how these advanced models are being used to enhance predictions about patient survival, recovery, and complications in ICU settings.
How is generative AI revolutionizing ICU outcome prediction?Traditionally, outcome prediction in intensive care units relied on clinical scoring systems, which, despite their utility, often lacked the accuracy needed for critical treatment decisions. The reviewed studies reveal that generative AI consistently outperforms traditional models, offering enhanced precision by processing vast amounts of patient data.
The authors categorized the applications into three primary use cases. The first involves data augmentation, where models generate synthetic data to balance datasets and address missing information, resulting in improved accuracy of downstream predictive models. For example, GAN-based techniques successfully compensated for class imbalances and improved mortality prediction models, achieving AUROC scores above 0.90 in multiple studies.
The second use case focuses on feature generation from unstructured data. Generative models extract meaningful parameters from sources like medical notes, radiology reports, and time-series data, which are then used by predictive models to improve prognostication. This integration of unstructured and structured data allows AI systems to detect subtle patterns that might be overlooked by clinicians.
The third use case showcases direct outcome prediction by generative models. Unlike traditional AI systems that serve as feature enhancers for predictive models, these generative frameworks take on the prediction task themselves. Large language models such as GPT-4 and specialized ICU-focused LLMs have demonstrated strong capabilities in predicting 30-day mortality, delirium, and neurological outcomes after cardiac arrest, with results that rival or surpass clinician-derived scores.
Which technologies are leading the shift in critical care forecasting?The study identifies two dominant generative technologies driving progress: Generative Adversarial Networks (GANs) and Generative Pretrained Transformers (GPTs). GANs have been widely used for generating synthetic medical data, balancing datasets, and imputing missing values, while GPT models have excelled in interpreting complex medical texts and generating predictive insights.
Other models, including Variational Autoencoders (VAEs) and hybrid approaches combining transformers with traditional AI, have also made significant contributions. For example, models that fused BERT-based embeddings with GAN-generated data improved predictions of ICU mortality by leveraging both textual and numeric data.
The dataset most commonly employed across the studies was the Medical Information Mart for Intensive Care (MIMIC), reflecting its importance as a benchmark resource for ICU-focused AI research. The frequent use of this dataset underlines the need for more diverse data sources to improve generalizability.
Despite impressive results, the review notes that most research to date has concentrated on short-term outcomes, such as in-hospital or 30-day mortality. The lack of studies predicting long-term outcomes like functional recovery, quality of life, or post-discharge complications highlights an important research gap.
What are the implications and challenges of using generative AI in ICU care?The findings underscore that generative AI has immense potential to transform ICU care by enabling faster, more accurate prognoses. This capability can directly impact treatment decisions, resource allocation, and patient counseling. However, several challenges must be addressed before these tools can be widely adopted.
One major concern is data privacy and security, particularly because generative models require extensive data to function effectively. Moreover, the computational intensity of these models raises questions about scalability and access in resource-limited settings. The study also warns of bias in models trained on unrepresentative datasets and highlights the risks associated with AI-generated errors or "hallucinations" in clinical predictions.
The ethical dimension is equally significant. Generative AI applications must balance technological efficiency with human oversight, ensuring that predictions support rather than replace clinical judgment. The authors emphasize that despite AI's advanced capabilities, the final treatment decisions must remain under the control of experienced medical professionals.
Going ahead, the review advocates for interpretability and explainability in generative AI models. Without clear insights into how these systems arrive at their predictions, clinicians may be hesitant to rely on them. Additionally, integrating AI models with electronic health record systems and obtaining regulatory approval are essential steps for real-world deployment.
14 Transformative AI Applications In Medicine - Psychology Today
A lot of people are worried about Artificial Intelligence (AI). What is AI going to do to our society? To our workforce? Is it going to make our jobs disappear? Is it going to harm us?
Personally, I am not worried about AI in medicine. Because I am a physician, I am more interested in what AI can do to help the medical profession. AI is becoming a transformative medical professional tool that still needs to be supervised by humans, but that is rapidly revolutionizing our medical future for the better, enhancing human medical capabilities.
Among many others, here are 14 game-changing medical applications of AI:
1. AI can analyze our tone of voice, speed of speech, and facial expressions and diagnose early depression, anxiety, or brain neuronal degeneration, which will often be missed by clinicians. That allows for more accurate diagnoses and earlier treatments for better longevity. [1]
2. AI can evaluate subtle color changes of our skin through a video of our face to extract our heart rate and blood oxygen saturation level, a technique called remote photoplethysmography (rPPG). [2]
AI can also study pictures of our eyes looking up, down, left, and right to detect COVID infection and several other diseases. [3]
3. During a routine examination with an EKG-enabled stethoscope, AI can diagnose early signs of heart failure that would not be detected by a human.[4]
4. Generative AI can already suggest diagnoses and possible treatments across all medical specialties based on the patient's clinical symptoms, laboratory, and radiology results. With large amounts of data from large populations of patients, we can predict which treatments for depression, anxiety, PTSD, cancer, and other diseases will work and which treatments will not work. [5]
This saves time and suffering, avoiding the side effects of the wrong medications.
For cancers, AI already helps physicians understand the genomics and proteomics of tumors. Based on those, AI recommends the best personalized course of action that can be used to fight against each individual cancer instead of using the one-size-fits-all model we used to have, and in some cases, that we still currently have. [6]
5. AI is already used for training new residents and surgeons with medical simulations. Medical and surgical students won't need to practice only on real bodies anymore.
6. AI can predict epidemic outbreaks. It can help us understand where the next outbreak will happen and where it's going to be most severe. [7]
7. Generative AI can help drug companies synthesize new drugs and can find clinical trial candidates for those new drugs. [8]
8. Security Monitoring of health care computers: We already use AI programs for cyber-protection with hospital monitoring systems 24 hours per day. AI can monitor all computers for all out-of-the-ordinary activity. If the AI sees something unusual, increased frequency, increased activity, or suspicious activity in a particular server, it will immediately shut down that server and send a message to the cybersecurity team so that they can start investigating.
9. Radiology: AI will be able to tell from a chest X-Ray, not only what is abnormal in the lungs, but also what else is abnormal in the surrounding tissues: bones, heart, vessels, lymph nodes, breasts, etc.
AI can tell better than a human if, in a CT scan, a nodule less than 1 cm in diameter is cancerous, and if the nodule is cancerous, whether there is a risk of recurrence after resection. [9]
In some facilities, AI is already used to produce radiology reports that are better understood by patients (who often see their reports before their physician calls them).
10. Pathology: AI will be used for analyzing cells and detecting cancerous cells in biopsies. It is especially useful in interpreting the meaning of borderline cells. Are the cells cancerous or not? AI can make the diagnosis better than a human in borderline cases because the computer (digital pathology) can see malignant elements of the biopsy that are so tiny that the human eye cannot discern them —even with a powerful microscope — and that would have been considered as normal before. [10]
11. Preventative Medicine: AI will help in the study of genetics and, based on the patient's individual genomic study, recommend personalized preventative actions for each individual.
12. In remote areas where the medical care is sparse and in areas where physicians and nurses are overwhelmed, it will be possible to have hospitals with AI-driven robots of physicians and nurses, AI-interpreted vitals, EKGs, X-Rays, blood tests, and robots administering treatments. Already, researchers at Tsinghua University in China have announced they have created a model of an AI hospital with virtual AI physicians who can treat 10,000 virtual patients in just a few days. [11]
13. Here in the US, AI is used in some medical offices to transcribe summary notes of office visits so that physicians can spend most of their time directly interacting with their patients.
14. Also, AI can summarize very complicated, long charts in a simple, short way that will save physicians time.
In summary, thanks to supervised-by-human Artificial Intelligence, the tasks of physicians will become simpler, faster, and more accurate. AI is a game changer that will allow better personalized medicine, better targeted treatments, and more satisfied patients.
References
[1] Zafar F, Fakhare Alam L, Vivas RR, Wang J, Whei SJ, Mehmood S, Sadeghzadegan A, Lakkimsetti M, Nazir Z. The Role of Artificial Intelligence in Identifying Depression and Anxiety: A Comprehensive Literature Review. Cureus. 2024 Mar 19;16(3):e56472. Doi: 10.7759/cureus.56472. PMID: 38638735; PMCID: PMC11025697.
[2] Chen W, Yi Z, Lim LJR, Lim RQR, Zhang A, Qian Z, Huang J, He J, Liu B. Deep learning and remote photoplethysmography powered advancements in contactless physiological measurement. Front Bioeng Biotechnol. 2024 Jul 17;12:1420100. Doi: 10.3389/fbioe.2024.1420100. PMID: 39104628; PMCID: PMC11298756.
[3] A New Screening Method for COVID-19 based on Ocular Feature Recognition by Machine Learning ToolsYanwei Fu1, Feng Li3, Wenxuan Wang2, Haicheng Tang3, Xuelin Qian2, Mengwei Gu1,4, Xiangyang Xue1,2 , 2020 https://www.Medrxiv.Org/content/10.1101/2020.09.03.20184226v5.Full.Pdf
[4] Point-of-care screening for heart failure with reduced ejection fraction using artificial intelligence during ECG-enabled stethoscope examination in London, UK: a prospective, observational, multicentre study, Bachtiger, Patrik et al. The Lancet Digital Health, Volume 4, Issue 2, e117 - e125
[5] Prehm I.M. Arnold, Joost G.E. Janzing, Arjen Hommersom, Machine learning for antidepressant treatment selection in depression, Drug Discovery Today, Volume 29, Issue 8, 2024,104068, ISSN 1359-6446,https://doi.Org/10.1016/j.Drudis.2024.104068 (https://www.Sciencedirect.Com/science/article/pii/S1359644624001934)
[6] LORIS robustly predicts patient outcomes with immune checkpoint blockade therapy using common clinical, pathologic and genomic features. Chang TG, Cao Y, Sfreddo HJ, Dhruba SR, Lee SH, Valero C, Yoo SK, Chowell D, Morris LGT, Ruppin E. Nat Cancer. 2024 Jun 3. Doi: 10.1038/s43018-024-00772-7. Online ahead of print. PMID: 38831056.
[7] Kraemer, M.U.G., Tsui, J.LH., Chang, S.Y. Et al. Artificial intelligence for modelling infectious disease epidemics. Nature 638, 623–635 (2025). Https://doi.Org/10.1038/s41586-024-08564-w
[8] Amit Gangwal, Antonio Lavecchia, Unleashing the power of generative AI in drug discovery,Drug Discovery Today, Volume 29, Issue 6, 2024, 103992, ISSN 1359-6446, https://doi.Org/10.1016/j.Drudis.2024.103992.(https://www.Sciencedirect.Com/science/article/pii/S135964462400117X)
[9] Effect of Human-AI Interaction on Detection of Malignant Lung Nodules on Chest Radiographs Jong Hyuk Lee, Hyunsook Hong, Gunhee Nam, Eui Jin Hwang, and Chang Min Park. Radiology 2023 307:5
[10] McGenity C, Clarke EL, Jennings C, Matthews G, Cartlidge C, Freduah-Agyemang H, Stocken DD, Treanor D. Artificial intelligence in digital pathology: a systematic review and meta-analysis of diagnostic test accuracy. NPJ Digit Med. 2024 May 4;7(1):114. Doi: 10.1038/s41746-024-01106-8. PMID: 38704465; PMCID: PMC11069583.
[11] https://www.Roboticsandautomationmagazine.Co.Uk/news/healthcare/worlds-first-ai-hospital-with-virtual-doctors-opens-in-china.Html
More referencesHow AI Is Transforming Healthcare In 2025
Ashish Sukhadeve, Founder and CEO of Analytics Insight, providing organizations with strategic insights on disruptive technologies.
In terms of its transformative effect, AI is doing for healthcare what electricity did for the industry. According to Fortune Business Insights, the global AI healthcare market was valued at $29.01 billion in 2024, and it's projected to grow to $504.17 billion by 2032. This represents a massive CAGR of 44%.
For healthcare leaders, the time to act is now; AI is rapidly advancing areas such as illness prevention, diagnostics and drug discovery, and it has the potential to significantly reduce administrative burdens. Here are just some of the areas to follow when it comes to AI in the sector:
Preventing IllnessI think one of AI's most valuable contributions is its ability to prevent illness. There are many wearables, AI health apps and the Internet of Medical Things (IoMT) available now. These consumer applications enable individuals to take charge of their own health. They allow users to monitor everything from heart rates to sleep cycles, turning passive patients into proactive participants.
Doctors also benefit. AI helps them understand a patient's daily habits, risks and medical history. This improves their ability to offer personalized care. I see prevention, Diagnosis
AI also excels in medical imaging and diagnostics. At Massachusetts General Hospital and MIT, AI detected lung nodules with 94% accuracy, outperforming radiologists, who achieved 65%. Similar results have been found for breast cancer detection. According to the Harvard School of Public Health, AI in diagnostics may reduce treatment costs by 50% and improve outcomes by 40%. AI is not just reducing medical errors; it is reducing medical anxiety.
I believe with AI improving both accuracy and outcomes, healthcare leaders must treat diagnostics not as a back-office upgrade, but as a frontline opportunity for transformation.
Drug DiscoveryBringing a drug to market can take over 10 years and cost $172 million on average. Only 13.8% of compounds make it to approval. AI can help reduce this timeline dramatically.
Systems like IBM Watson can process vast amounts of medical journals and case studies. DeepMind uses neural networks to solve complex health problems with learning algorithms. AstraZeneca's AI, trained on data from 500,000 people, can now predict diseases like Alzheimer's before symptoms appear. These tools can be used not only to improve timelines but also to save countless lives.
Caring For The Aging PopulationAs populations age, the demand for elder care rises. AI-powered robots are easing this burden, even when it comes to things like laundry. They are increasingly used to assist in physical therapy, provide companionship and reduce hospital visits. These machines can now hold conversations that stimulate mental engagement in older adults.
The lesson is clear: Embracing AI not only addresses labor shortages but also opens new paths to compassionate, scalable elder care.
Administration SupportAI is improving what patients don't see, the administration. Microsoft's Dragon Copilot can create real-time notes during clinical consultations. Germany's Elea AI reports that it can cut testing and diagnosis from "weeks to hours." These and many other tools are giving doctors more time focusing on patient care.
These tools are also helping save time when it comes to clinical decision making, and as evidenced by the Alzheimer's prediction example, AI increasingly supports early warnings. It can now identify high-risk patients using data patterns, including genetic and lifestyle factors. Tools like ChatRWD are replacing and outperforming generic AI models.
ConclusionAI is not replacing doctors but empowering them. It now impacts every layer of healthcare, from diagnosis to prevention, from robots to research. Patients can be better informed. Doctors are better equipped. Systems are more efficient.
In 2025, I believe the most important medical tool is not a scalpel; it's data, and AI is the surgeon. The future of healthcare is not waiting on some distant horizon; it's already here.
Comments
Post a Comment