Generative artificial intelligence (GAI), such as the one behind OpenAI’s DALL-E and ChatGPT, is expected to play a significant role in advancing the field of oncology, according to Daphne Koller, an AI expert and CEO of Insitro.
Koller spoke at a workshop organized by Stanford University’s Human-Centered AI institute, highlighting the use of GAI to interpret histopathology language and identify potential drug targets. Using machine learning, the AI technology is able to analyze histology images generated from cancerous tissue, providing insights that are not readily apparent to human pathologists.
Additionally, the Insitro team used GAI to create synthetic tissue images, greatly expanding the sample size and enabling deeper analysis. By leveraging these images along with a genetic assay, the team identified novel candidate drug targets for conditions such as triple-negative breast cancer.
Koller emphasized the potential of GAI in biology, describing it as a way to understand complex and subtle data patterns that are beyond the capabilities of the human brain. She also highlighted the merging of data/machine learning/AI with quantitative biology to create digital biology, which has broad implications for health, environment, and other fields.
Koller’s insights were part of a broader discussion on the use of GAI across different data types and its potential for future advancements in medicine and other disciplines. The workshop showcased various applications of multi-modality GAI, pointing to its growing importance in research and innovation.
