Photo by Google Research
“The future of healthcare lies in the hands of artificial intelligence.”
Google Research is no stranger to tackling complex challenges, from quantum computing to flood forecasting. However, one area has particularly captured their attention – healthcare. They believe AI can catalyze the creation of more personalized, accessible, and practical solutions in this domain.
A Journey Towards AI-Driven Healthcare
Over the years, Google Research has collaborated closely with healthcare providers, researchers, and industry partners to publish research, create open-source tools, and build AI systems with the potential to improve health outcomes globally. This has included helping radiologists detect breast cancer and supporting diabetic retinopathy screening.
One of their significant projects was developing a large language model (LLM) called Med-PaLM, fine-tuned for medicine. It achieved state-of-the-art performance in answering U.S. Medical License Exam-style questions. However, their research continued. Their goal is to positively impact healthcare responsibly and continue exploring the potential of generative AI in medicine.
Exploring the Potential of Generative AI in Medicine
In their continuous effort to improve these tools’ safety, accuracy, and equity, Google Research has developed ELIXR, a novel approach to integrating LLMs with medical imaging models to create specialized multimodal models. Additionally, they are having clinicians evaluate Med-PaLM’s performance and publicly share their results. They are partnering through Google Cloud to explore how these medically fine-tuned models can transform healthcare.
Bringing Research to Reality
Google Research recognizes the importance of clinical validation and application. For this reason, they are partnering with healthcare organizations to validate their research in natural clinical settings. They are working with Jacaranda Health in Kenya to improve their fetal ultrasound AI models, with Osaka University to improve their dermatology classifiers, and with a group of healthcare organizations to understand where and how Med-PaLM 2 technology is most helpful.
They are also partnering with PacBio, a developer of genome sequencing instruments, to improve genomic analysis using DeepVariant and DeepConsensus. These partnerships are essential to scaling their work and improving the health of millions globally.
Sharing Knowledge and Empowering Innovators
In their journey towards impact, Google Research has learned a lot about the unique challenges of integrating AI-powered solutions into healthcare systems. One of their lessons is building more equitable datasets and identifying ways to embed their solutions into clinical workflows effectively and safely.
Beyond sharing these learnings, they make their tools publicly available to health researchers and innovators. Through open-sourcing technologies like Open Health Stack and CXR Foundation, they are empowering others to build next-generation digital health solutions that can save time and improve health outcomes.
Through both foundational and applied research in health AI, Google Research hopes to help people envision a future in healthcare that is more accessible, accurate, and equitable.