During his HLTH 2024 keynote session Byte-Sized Care: AI's Mega Impact on Health, Dr. Taha Kass-Hout detailed how artificial intelligence and cloud computing will transform healthcare. Kass-Hout also spoke about the different ways AI will open up access to quality healthcare for more than 4.5 billion people in the world currently not covered by essential health services.
As the Global Chief Science and Technology Officer at GE HealthCare, Kass-Hout is uniquely positioned to analyze the technological trends that will push at the boundaries of what’s possible in healthcare. Prior to GE HealthCare, Kass-Hout led a wide variety of health AI initiatives at Amazon Web Services. He also served in key leadership roles under President Obama, where he was the first Chief Health Informatics Officer at the FDA.
In case you missed Dr. Kass-Hout’s keynote, here are seven key takeaways that will get you up to speed on the latest trends shaping the adoption of technology in healthcare:
1. AI adoption in healthcare remains slow despite the explosion of data
By 2025, the world is expected to generate more than 180 zettabytes of data, with more than one-third originating from healthcare. Yet, only half of the healthcare industry has adopted cloud services to manage this data effectively. The healthcare sector's hesitation in embracing AI contrasts with industries like music and e-commerce, where AI has proven effective in making sense of large datasets. 
2. Multimodal data and custom AI models pose challenges
Most of the healthcare industry’s data is unstructured (e.g., medical images, notes, audio recordings, device readings), making it difficult for traditional analytics and most machine learning algorithms to process. The multimodal nature of the data has proven to be a barrier to AI adoption in healthcare.
Additionally, traditional approaches require vast amounts of domain-specific data and manual feature engineering for different disease states, leading to costly and resource-intensive development cycles—further underscoring why current approaches are not sustainable.
3. Foundation models offer a scalable solution
Foundation models can help mitigate the challenges presented by the need for constant customization. These models can be trained on a broad range of multimodal data enabling them to be applied across different healthcare tasks without extensive retraining. GE HealthCare has been pioneering the development of foundation models in healthcare. To give one example, SonoSAMTrack is a foundation model in the research stage that segments objects of interest on ultrasound images. It adapts to new data and tasks without requiring extensive retraining.
In addition, the GE HealthCare research and development team has built the industry’s first full-body X-ray foundation model. Even though our model is a full-body foundation model – in our internal testing, it outperforms even the most specialized models.
Read now: Introducing a Multimodal AI HealthCare Model Trained on Comprehensive New Dataset
4. Clinicians face cognitive overload and workflow inefficiencies
According to the World Health Organization, by 2050, the global population of older people is expected to double, to 2.1 billion. As the global population increases, so does the exponential growth of data, spanning industries such as genomics, ingestibles and implantables, wearables, sensors, social media, clinical data, and electronic health records.
In addition to the explosion of data, clinicians struggle with workflow inefficiencies. Managing vast amounts of information from numerous sources leads to delays, errors, and burnout. Providing adequate resources and access to care is essential, but orchestrating the necessary coordination and workflows across numerous touchpoints and overwhelming data is challenging. This complexity contributes to delays in care, medical errors, and clinician burnout. Nearly half of clinicians report high levels of burnout, and according to a 2024 NSI National healthcare retention and RN staffing report, hospitals average 100% staff turnover every five years.
5. CareIntellect enhances the patient care journey and workflow efficiencies
CareIntellect is GE HealthCare’s new offering of clinical and operational applications designed to optimize care delivery for across disease states. Built on a cloud-first digital foundation, CareIntellect is designed to help customers optimize care delivery and quality across health networks and disease states.
Once a customer installs their first CareIntellect application, they would be able to activate additional applications (which will become available in the future) quickly, securely, and without costly integration. All applications will leverage a common data layer and integrate with a provider’s single sign-on application, enabling clinicians to focus on their patients.GE HealthCare is developing multiple applications in the CareIntellect offering, focusing on both clinical and operational use cases. Our first application is CareIntellect for Oncology that is designed for oncologists and care teams. The application aggregates and summarizes multi-modal patient data from disparate systems, using generative AI to summarize patient notes and reports
Learn more about CareIntellect for Oncology
6. Project Health Companion is emblematic of the era of agentic AI
The Health Companion project, currently in research, represents a new era in AI: a shift from models generating text, images, and video to agents that can interact with their environment and act to accomplish predefined goals.
The Health Companion project explores whether an agentic AI approach driven by multiple agents – each an expert in a particular area (e.g., genomics, radiology, pathology, etc.) - could help physicians streamline their clinical decision-making and deliver more personalized care. The project’s aim will be to use agentic AI to analyze multi-modal data and generate treatment plan recommendations.
7. AI will unlock the path to equitable healthcare
AI has the potential to bridge healthcare disparities and help enable access to quality care for the 4.5 billion people in the world currently uncovered by essential services.
To give just one example, Vscan Air is a portable ultrasound device that enables patients to receive quality imaging from the comfort of their homes, wherever they are located. Powered by Caption AI software, it provides real-time guidance to non-expert healthcare users, allowing even those without extensive training to capture diagnostic-quality ultrasound images.
Every breakthrough in AI for healthcare promises quality care for everyone. The future relies on collaboration between human expertise and artificial intelligence to transform possibilities into reality.
Learn more about the innovations that GE HealthCare showcased at HLTH 2024 and sign up to stay up to data on how the convergence of AI and cloud computing is shaping the next generation of healthcare.
