In a letter addressed to Google CEO Sundar Pichai, United States Senator Mark R. Warner (D-VA), Chairman of the Senate Select Committee on Intelligence, expressed concerns regarding the rollout of Med-PaLM 2, a large language model (LLM) designed for use in healthcare settings.
Google announced in April that selected healthcare clients will have restricted access to Med-PaLM 2 to evaluate the model and offer comments on its potential healthcare use cases. The tech giant's release coincided with a peak in generative AI interest, according to Health IT Analytics.
However, numerous doctors and academics have called for a cautious and rigorous approach to introducing AI into medicinal settings.
Calls For Greater Transparency
The letter from Senator Mark Warner highlights some of the underlying worries about the use of generative AI in healthcare. The legislator strongly emphasizes the need for increased openness, protecting patient privacy, and creating ethical standards.
Senator Warner acknowledges the immense potential of AI to improve patient care and health outcomes. However, he worries that premature deployment of untested technology could erode trust in medical professionals and institutions, worsen racial disparities in health outcomes, and increase diagnostic and care-delivery errors.
Amid concerning reports that Google is testing an AI chatbot in real medical settings, I’m calling for more transparency and accountability. We must put patient wellbeing over risky rollouts.
— Mark Warner (@MarkWarner) August 8, 2023
One of Senator Mark Warner's worries is the possibility of errors in the Med-PaLM 2 model and the moral issues surrounding its use, data sources, and possible biases. The lawmaker expressed concern about Google's efforts to avoid bias and let patients know about AI's role in their treatment.
In his letter, Senator Warner also takes issue with current efforts in healthcare that use generative AI. He makes the case that major tech firms like Google and Microsoft may be racing to create and implement healthcare AI models to attain market domination, possibly endangering patient privacy and safety.
The legislator also asks Google twelve direct questions about anything from the data sources used to train the AI model to patients' capacity to refuse AI participation in their treatment.
Large Number of American Adults Not At Ease with Healthcare AI
The World Health Organization (WHO) has also advised caution when using AI-generated huge language model tools to safeguard human welfare, security, and autonomy, as per a report from The Messenger.
Meanwhile, Forbes reported that, according to a Morning Consult poll, around 70% of American citizens have voiced reservations regarding the use of AI in healthcare, with Baby Boomers having the highest reservations (77%).
The survey indicated that respondents feel most at ease having AI take notes during medical sessions and aid with administrative work. However, the use of AI in surgery was correlated with the most minor level of comfort.
