Generative AI and large language models, such as the ones used in ChatGPT, are already being utilized in various areas of healthcare. Developers are working to create time-saving tools and providers are using this technology for patient engagement and clinical decision support. However, there are still questions about whether generative AI is ready for widespread use and what it can truly accomplish at this early stage. Vrinda Khurjekar, senior director of Americas business at Searce, an AI consultancy, provides expert answers to these questions and more.
According to Khurjekar, generative AI is the next major technological advancement that will transform our work and daily lives, following the revolution brought about by the internet. Its applicability across industries is already evident, including in healthcare. Implementations such as conversational appointment-setting processes and summarizations for doctors to review relevant information are significantly impacting patient experience and overall healthcare efficiency. While there are still challenges to overcome, such as minimizing errors and defining data privacy and sharing rules, the results are promising. With proper governance and controls, generative AI can be reliably used for a wide range of use cases in the future.
However, there are challenges for generative AI to achieve widespread use. The lack of clarity in where to apply the technology leads to non-value-adding use cases being pursued. Organizations need to clearly identify the purpose and benefits of implementing generative AI. Additionally, many models currently experience hallucinations and need further refinement for stable functionality. The high cost of implementing and the scarcity of talent also hinder mainstream adoption. Until economies of scale are achieved, access to generative AI will be limited to organizations that can invest in it.
For executives and clinicians in hospitals and health systems faced with new generative AI tools, Khurjekar advises establishing checks for vendors’ security, compliance, and governance policies. Internal guidelines may need to be revised to accommodate the nuances of AI technologies. A zero-trust policy, stricter compliance audits, and investment in internal experts are crucial for successful implementation.
Regarding Oracle’s Clinical Digital Assistant, Khurjekar finds it exciting to see generative AI services being introduced in the healthcare industry. The tool aims to streamline administrative tasks through voice commands, which can revolutionize clinical visits. While the vision is promising, the practical rollout and features are still to be seen within the next 8-12 months.
Looking five years into the future, Khurjekar expects generative AI to become mainstream in healthcare, leading to revamped business processes. Niche technology players will emerge, and compliance and regulatory guidelines will undergo a complete overhaul to ensure accountability and privacy. Preventive care will also become more prevalent, improving individual well-being and life expectancy.
In summary, generative AI holds immense potential for healthcare, but there are challenges to overcome before widespread use. Proper governance, clarity in use cases, and investment in internal expertise are essential for successful implementation.
