# Large Language Models * **Definition:** Advanced AI systems designed to understand and generate human-like text, often trained on vast amounts of text data, and used in various applications including healthcare. * **Taxonomy:** CTO Topics / Large Language Models ## News * Selected news on the topic of **Large Language Models**, for healthcare technology leaders * 16.3K news items are in the system for this topic * Posts have been filtered for tech and healthcare-related keywords | Date | Title | Source | | --- | --- | --- | | 5/22/2025 | [**Anthropic, now worth $61 billion, unveils its most powerful AI models yet-and they show ... - Fortune**](https://fortune.com/2025/05/22/anthropic-new-models-ai-openai-google/) | [[Fortune]] | | 5/22/2025 | [**Lloyds and Nationwide to use UK finance sector LLM - Computer Weekly**](https://www.computerweekly.com/news/366624472/Lloyds-and-Nationwide-to-use-UK-finance-sector-LLM) | [[Computer Weekly]] | | 5/12/2025 | [**Revolutionizing Healthcare: The Synergy of Knowledge Graphs and Large Language Models**](https://www.healthcareittoday.com/2025/05/12/revolutionizing-healthcare-the-synergy-of-knowledge-graphs-and-large-language-models/) | [[Healthcare IT Today]] | | 5/8/2025 | [**Early days for small language models and AI at the edge - Computer Weekly**](https://www.computerweekly.com/feature/Early-days-for-small-language-models-and-AI-at-the-edge) | [[Computer Weekly]] | | 4/28/2025 | [**The Role of AI in Delivering Personalized Nutrition for Patients With Cancer: Julia Logan, BS**](https://www.ajmc.com/view/the-role-of-ai-in-delivering-personalized-nutrition-for-patients-with-cancer-julia-logan-bs) | [[AJMC]] | | 4/27/2025 | [**Large Language Models Could Help Bridge Gaps in Cancer Nutrition Care: Julia Logan, BS**](https://www.ajmc.com/view/large-language-models-could-help-bridge-gaps-in-cancer-nutrition-care-julia-logan-bs) | [[AJMC]] | | 4/19/2025 | [**AMD's CTO says AI inference will move out of data centers and increasingly to phones and laptops**](https://www.businessinsider.com/ai-workloads-transition-inference-amd-mark-papermaster-edge-devices-2025-4) | [[Business Insider]] | | 3/4/2025 | [**LlamaIndex launches a cloud service for building unstructed data agents**](https://techcrunch.com/2025/03/04/llamaindex-launches-a-cloud-service-for-building-unstructed-data-agents/) | [[TechCrunch]] | | 2/24/2025 | [**Transforming Clinical Data for Point-of-Care Intelligence**](https://www.healthitanswers.net/transforming-clinical-data-for-point-of-care-intelligence/) | [[Health IT Answers]] | | 2/21/2025 | [**Roundup: AI and cloud tackle cyber risk and improve workflows - Healthcare IT News**](https://www.healthcareitnews.com/news/roundup-ai-and-cloud-tackle-cyber-risk-and-improve-workflows) | [[Healthcare IT News]] | | 2/1/2025 | [**Weekly Roundup - February 1, 2025**](https://www.healthcareittoday.com/2025/02/01/weekly-roundup-february-1-2025/) | [[Healthcare IT Today]] | | 1/30/2025 | [**AI Large Language Models Market Soars at 79.8AGR - Valuates Reports - PR Newswire**](https://www.prnewswire.com/news-releases/ai-large-language-models-market-soars-at-79-8-cagr--demand-for-chatbots-content-generation--nlp-rises--valuates-reports-302364522.html) | [[PR Newswire]] | | 1/29/2025 | [**LLMs in Healthcare: A Measured Path to Impact**](https://www.healthcareittoday.com/2025/01/29/llms-in-healthcare-a-measured-path-to-impact/) | [[Healthcare IT Today]] | | 1/28/2025 | [**Technology and trends shaping healthcare in 2025 - Becker's Hospital Review**](https://www.beckershospitalreview.com/healthcare-information-technology/technology-and-trends-shaping-healthcare-in-2025.html) | [[Beckers Hospital Review]] | | 1/2/2025 | [**What Analysts Think of Nvidia Stock Ahead of CEO Jensen Huang's CES Keynote**](https://finance.yahoo.com/news/analysts-think-nvidia-stock-ahead-164723741.html) | [[Yahoo Finance]] | | 1/2/2025 | [**Alibaba Leverages Scale and Deep Pockets to Slash AI LLM Prices To Gain Market Share**](https://www.aol.com/alibaba-leverages-scale-deep-pockets-141627154.html) | [[AOL]] | | 12/29/2024 | [**Analyst Explains New AI Growth Catalysts for Amazon.com (AMZN) - Yahoo Finance**](https://finance.yahoo.com/news/analyst-explains-ai-growth-catalysts-201136606.html) | [[Yahoo Finance]] | | 12/20/2024 | [**Age against the machine-susceptibility of large language models to cognitive impairment: cross sectional analysis**](http://www.bmj.com/content/387/bmj-2024-081948.short?rss=1) | [[BMJ]] | | 12/17/2024 | [**Ryght AI Launches Global Clinical Trial Site Network with the University of Southern ...**](https://www.prnewswire.com/news-releases/ryght-ai-launches-global-clinical-trial-site-network-with-the-university-of-southern-california-as-their-first-academic-site-302330828.html) | [[PR Newswire]] | | 11/27/2024 | [**Artificial Intelligence - Healthcare IT News**](https://www.healthcareitnews.com/taxonomy/term/7341/m89gsv6dzcjz.jsp/page/222?type=video) | [[Healthcare IT News]] | | 10/16/2024 | [**FDA strengthens AI regulation to ensure patient safety and innovation in healthcare**](https://www.news-medical.net/news/20241016/FDA-strengthens-AI-regulation-to-ensure-patient-safety-and-innovation-in-healthcare.aspx) | [[News Medical Net]] | | 8/8/2024 | [**Novel AI model may enhance health data interoperability - Medical Xpress**](https://medicalxpress.com/news/2024-08-ai-health-interoperability.html) | [[MedicalXpress]] | | 7/17/2024 | [**What Will Healthcare AI Developers Do When and If the Data Well Runs Dry? - AHIMA**](https://www.ahima.org/education-events/artificial-intelligence/what-will-healthcare-ai-developers-do-when-and-if-the-data-well-runs-dry/) | ahima.org | | 6/27/2024 | [**Healthcare Analytics Information, News and Tips - TechTarget**](https://www.techtarget.com/healthtechanalytics/) | techtarget.com | | 1/10/2024 | [**Generative artificial intelligence models effectively highlight social determinants of health in doctors' notes**](https://www.msn.com/en-us/health/other/generative-artificial-intelligence-models-effectively-highlight-social-determinants-of-health-in-doctors-notes/ar-AA1mNIaF) | msn.com | ## Topic Overview (Some LLM-derived content — please confirm with above primary sources) ### Key Players - **OpenAI**: A leading AI research organization known for developing large language models like ChatGPT. - **Google**: A major tech company that has developed large language models such as PaLM, contributing to advancements in AI. - **Northwestern Medicine**: Developed FHIR-GPT, a model to enhance interoperability of electronic health records using large language models. - **NVIDIA**: A key player in AI hardware and software, providing the infrastructure necessary for training and deploying large language models. - **Nym**: A company utilizing large language models for specific healthcare tasks, achieving high accuracy in medical coding and documentation. - **Meta**: The parent company of Facebook, which has developed Llama, a large language model aimed at various applications. - **Microsoft**: A technology giant that partners with various companies to integrate AI capabilities, including large language models, into their products. - **Eleos**: A leader in AI for behavioral health, utilizing large language models to improve care outcomes and enhance clinical documentation. - **Dynex**: A technology company collaborating with InCor to leverage large language models for healthcare improvements. - **FDA**: The U.S. Food and Drug Administration, which is responsible for regulating AI technologies in healthcare, including large language models. - **Oracle**: Offers cloud services that facilitate the use of generative AI and large language models in enterprise applications. - **Anthropic**: An AI firm that has partnered with Amazon to integrate its large language models into Amazon's Bedrock platform. ### Partnerships and Collaborations - **Cognizant and Google Cloud**: Collaborating to streamline administrative tasks using large language models. - **Amazon and Anthropic**: Partnered to integrate large language models into Amazon's Bedrock platform for generative AI applications. - **ConcertAI and NVIDIA**: Collaboration to optimize technologies for oncology clinical development, utilizing large language models for data processing. - **University of California San Diego**: Researchers are utilizing large language models to automate functional genomics research, showcasing the integration of AI in medical research. - **Dynex and InCor**: Collaboration to enhance healthcare processes using neuromorphic quantum computing and large language models. - **Lockheed Martin and IBM**: Integrating IBM's Granite LLMs into Lockheed Martin's AI Factory tools to enhance development capabilities. - **NASA and IBM**: Developed INDUS models for scientific domains, enhancing performance in biomedical tasks. - **Linklaters and Migrasia**: Developed an AI-powered chatbot for migrant workers, showcasing the application of LLMs in legal support. - **Stellar and Healthcare Organizations**: Stellar collaborates with mid-sized healthcare organizations to integrate generative AI and LLMs into their operations. - **Memorial Sloan Kettering and AWS**: Utilizing AI tools to analyze clinical, imaging, and genomic data for cancer research. - **Litify and AWS**: Collaborating to develop responsible AI solutions for the legal industry, focusing on data security. - **Alation and Salesforce**: Collaborated to improve data governance and lineage within Salesforce Data Cloud, supporting AI initiatives. - **LG and Samsung with Microsoft**: Partnered to integrate Microsoft's Copilot AI into their smart TVs, enhancing user experience with AI functionalities. - **Squirro and Synaptica**: Squirro acquired Synaptica to integrate semantic graph technology with generative AI for improved knowledge management. - **Deloitte and NVIDIA**: Launched AI Factory as a Service to enhance workload management and optimize workforce efficiency. - **Every Cure and Google Cloud**: Using AI for drug repurposing to improve patient outcomes. - **Trustworthy and Responsible AI Network (TRAIN)**: Launched to promote responsible AI development in healthcare, including rural healthcare providers. - **Mayo Clinic and Abridge**: Implementing an AI-powered documentation platform to connect around 20,000 clinicians, reducing administrative burdens. - **Academy Health**: Announced new projects utilizing EHR and claims data to address healthcare issues, emphasizing the role of AI in data analysis. - **Switchboard, MD and ThreatAware**: Launching an AI-driven system to identify and prioritize potential infectious disease cases. ### Innovations, Trends, and Initiatives - **Large Language Models (LLMs)**: Being integrated into healthcare for tasks such as clinical documentation, patient engagement, and enhancing diagnostic accuracy. - **MedLM**: A large language model transforming diagnostics and patient interactions in healthcare by integrating extensive datasets. - **Generative AI Integration**: Large language models are increasingly being integrated into various tech services, enhancing user experiences in applications like search engines and photo editing tools. - **Combining LLMs with Traditional Methods**: Recent studies indicate that integrating large language models with traditional AI methods can enhance the accuracy of early cognitive decline detection. - **INDUS Models**: Specialized large language models trained on 60 billion tokens for scientific domains, improving biomedical task performance. - **Generative AI Growth**: The global generative AI market is projected to grow significantly, driven by the deployment of large language models. - **Healthcare AI Integration**: Large language models are being integrated into healthcare for tasks like patient communication and clinical decision support. - **LLMs in Administration**: Large language models are automating tasks like insurance claims processing and clinical documentation, reducing clinician burnout. - **BioChatter**: An open-source Python framework designed to lower barriers for biomedical researchers in deploying large language models. - **FHIR-GPT**: A large language model that converts EHR data into standardized health resources, promoting interoperability. - **Generative AI in Healthcare**: Generative AI and LLMs are being used to automate digital medical conversations, improving efficiency in patient care. - **AI in Diagnostics**: Models like Google's MedPaLM demonstrate high accuracy in medical examinations, showcasing AI's potential in diagnostics. - **LlamaIndex**: A platform enabling developers to create custom agents that work with unstructured data, enhancing AI applications in healthcare. - **FHIR-GPT Development**: A new AI model that enhances the conversion of EHR data into standardized formats, improving healthcare data interoperability. - **AI in Medical Device Innovation**: Enhancing design processes through AI technologies, including LLMs. - **o3-mini Model by OpenAI**: Designed to enhance accessibility to advanced AI with self-fact-checking capabilities. - **Mental Health Care Enhancement**: LLMs are being explored for improving diagnostics and treatment in mental health care by analyzing clinical notes. - **Ambient AI**: Utilizing LLM-powered ambient AI to alleviate physician burnout by streamlining clinical documentation and patient interactions. ### Challenges and Concerns - **Regulatory Challenges**: The unpredictable outputs of large language models pose challenges for clinical decision-making, necessitating ongoing scrutiny and validation. - **Bias in Data Representation**: Concerns about biases in training data affecting the performance and fairness of LLMs in healthcare applications. - **Data Security and Privacy**: Significant hurdles in implementing LLMs in healthcare, including interpretability and patient privacy. - **User Literacy Gaps**: Challenges in ensuring that healthcare professionals are adequately trained to use LLMs effectively. - **Recursive Data Generation Risks**: Research indicates that AI models may degrade when trained on data generated recursively from previous models, suggesting a cycle of diminishing performance. - **Data Quality Issues**: Concerns about the quality and completeness of unstructured healthcare data used in AI models, leading to biased analyses. - **Integration Costs**: High costs associated with training and integrating advanced AI models into healthcare systems. - **Data Privacy and Accuracy**: Concerns regarding the accuracy of LLMs and data privacy are significant as these technologies are integrated into healthcare settings. - **Data Privacy**: Concerns regarding the handling of sensitive patient data when utilizing LLMs in healthcare. - **Cognitive Limitations**: Studies indicate potential cognitive impairments in LLMs, raising concerns about their reliability in medical diagnostics. - **Performance Variability**: Challenges in ensuring consistent performance of AI systems in clinical settings. - **Data Security**: Concerns regarding the security of sensitive healthcare data when implementing generative AI. - **Ethical Considerations**: The need for guardrails to mitigate risks such as biased outputs and legal liabilities associated with LLMs. - **Data Privacy and Security**: Concerns regarding the use of closed LLMs from private companies, which may compromise data privacy and vendor dependence. - **Technical Costs**: High costs associated with deploying LLMs in healthcare settings can be a barrier to adoption. - **Data Management Issues**: Organizations face challenges in managing unstructured data, which is essential for effective AI implementation and security risk identification. - **Ethical and Legal Issues**: Integration of LLMs into EHRs raises concerns about unconsented data use, AI-related malpractice accountability, and the 'black box' nature of LLMs. - **AI Bias**: Addressing biases in AI models is essential to ensure ethical and effective healthcare applications. - **Bias and Transparency**: Ethical concerns about bias in AI models and the need for transparency in AI decision-making.