KT Curry MS, CGC (She/Her) ; Jeffrey Bissonnette, MSc, CGC (He/Him) ; Anna McGill, MS, CGC (She/Her) ; Elizabeth Wiley, MS, CGC (She/Her)
Breathtaking advances in computational techniques that harness human knowledge and apply it to tasks like artificial-reasoning promise to revolutionize many aspects of the human experience, including the practice of medicine. These techniques, collectively referred to as AI, are best exemplified in newly open-access language models like ChatGPT and BARD, which capitalize on vast training-data covering a wide range of subjects, and on statistical techniques applied to language and prediction that produce surprising and sometimes deceptively convincing results. Other more focused techniques aim to understand the nuances of genetic language and present the evidence-based results to human users. The potential power of these tools is apparent, and along with it comes trepidation regarding what should or should not be incorporated into clinical genetics.
Genetic Counselors (GCs) are in a unique position to usher in a new age of clinical and diagnostic genetics with the ability to leverage this technology’s full potential. Many of us find ourselves in a situation that was inconceivable during our training. As AI becomes increasingly utilized in science and medicine, it will be our challenge as GCs to embrace this technology and become leaders in the field. We appreciate the complexities of clinical genetics and the importance of the accuracy, standardization of methods, and ethical concerns that are associated with providing/interpreting genomic information.
Understandably, AI raises ethical concerns and challenges related to privacy, biases, transparency, and impact on the job market. GCs may fear that AI could replace their position. Patients may fear a relinquishing of privacy or an erosion of the human connection. Physicians may fear reduced accuracy of results. However, AI is merely a tool. Ensuring judicious application is more reasonable than abandoning this tool altogether. For instance, AI can be used to augment our everyday roles. The authors are among dozens of genetics professionals who work at a genomics intelligence company, using AI tools and AI-derived evidence to curate the clinical exome from published evidence. Use of this technology has promoted workflow efficiency, improved diagnostic rates, and allowed us as human experts to focus on the cognitive aspects of our duties rather than the data-gathering. As GCs, we leverage our clinical expertise in disease-phenotype knowledge, variant discrepancies, ethical considerations/applications, and technical limitations/strengths of functional assay results. This has bolstered our resiliency and dedication to finding accurate information for clinical, pharma, and research use. Rapid AND accurate review of these data would not be possible without assistance from our AI-capabilities.
Some GCs may gravitate to industry to impact whole cohorts of patients or whole categories of scientific methods, practices, policies, etc. AI-derived tools have been specialized to the application of genomics to aid in genetic nomenclature resolutions, transcript issues, published genomic dataset searches, and beyond. We ask ourselves every day, “What can AI do for us to increase efficiency?” so we humans can spend our time focusing on the art of genomic interpretation. Additional areas of opportunity in AI include more effective communication with patients/families, preparation of educational materials, optimizing workflows, and improving data gathering.
A benefit of our GC training is knowing the patient’s experience from the ground up. For GCs, it is ingrained that each patient, each test report, each variant of interest matters. The people behind the samples and datasets matter. Experiencing the impact on a one-to-one human scale in the clinical genetics space is instructive when thinking about examining big data. In this spirit, challenges in AI may include representing underserved communities, privacy, accuracy, patient consent, and sensitivity in the field of genetics. AI is only a tool to aid in specific tasks. As GCs, we will be the stewards of appropriate application of these tools in the service of better treating our patients.
GCs are well-suited to embrace the opportunities and address challenges the use of AI poses in the clinical genomics space and industry. We encourage you to identify a professional task you perform regularly and explore whether there is an AI-enabled tool that could increase efficiency and accuracy. Consider this a challenge! But also, an experiment. Be scientific and skeptical in your approach and consider the best way to implement the use of the new tool - not to replace entirely what you do, but rather to optimize.
KT Curry MS, CGC (She/Her) is a Quality Assurance Specialist in Genetic Variant Interpretation at Genomenon and is the current co-chair of the AI/ML subcommittee in the Genomic Technologies SIG. She enjoys providing genomic education to the medical community and nomenclature sleuthing.
Jeffrey Bissonnette, MSc, CGC (He/Him) is the Director of Genomic Curation at Genonenon. He is passionate about utilizing software solutions to make genomic information more widely accessible to patients and the medical community.
Anna McGill, MS, CGC (She/Her) is the Product Quality Manager at Genomenon. Anna has experience in clinical cancer genetic counseling and has spent the past decade specializing in variant interpretation. She is dedicated to ensuring that patients and clinicians have access to all available variant level data that is appropriately identified and properly categorized for the most accurate classification.
Elizabeth Wiley, MS, CGC (She/Her) is a Product Quality Specialist at Genomenon. She has spent most of her genetic counseling career working with hereditary cancer syndromes, both in the clinical and laboratory settings. She is passionate about accurate variant interpretation and is excited to be working on a team dedicated to making genomic information accessible to patients, clinicians, and researchers