The AI sector is concerned that as politicians and officials in Washington delve into the management of artificial intelligence in healthcare, there is a likelihood of mishandling the situation.
Bob Wachter, the head of the Department of Medicine at the University of California, San Francisco, describes this as an intricate issue, stating, “it’s an extremely difficult problem.” He also warns of the risk of overregulation and hasty interventions.
The impact of AI on healthcare is becoming increasingly significant, with 692 AI products having obtained FDA approval. These products streamline various tasks such as patient scheduling, staffing calculations in emergency departments, and documentation of clinical visits. Additionally, AI is now supporting pharmacists in interpreting diagnostic images and tests. Wachter even mentions consulting a specialized version of GPT-4, a substantial language model developed by OpenAI, for complex cases.
The government is playing catch-up in response to the evolving landscape shaped by AI and its anticipated transformations.
Michael Yang, a senior managing partner at OMERS Ventures, expressed in an email that politicians are notably behind the curve, contrasting with the substantial investments made by his peers in the industry. Reports from Rock Health reveal investments of nearly $28 billion in online health companies focusing on AI.
The evolving nature of AI poses a challenge for regulators, unlike traditional medicines with stable compositions over time. Efforts are underway to establish clear guidelines and safeguards as both the White House and various health-focused organizations collaborate on regulatory frameworks. Congressional attention is also evident, with the Senate Finance Committee holding a hearing on AI in healthcare on February 8.
The surge in lobbying activities accompanies the regulatory environment. CNBC reports a 185% increase in companies advocating for AI-related policies. To educate the public on the benefits of AI, the trade group TechNet has initiated a $25 million campaign that includes television advertisements.
Navigating the early stages of AI technology poses significant challenges in wise management, as highlighted by Bob Kocher, a partner at Venrock and former member of the Obama administration.
Lawmakers have engaged with Kocher on AI legislation, focusing on the potential hurdles in healthcare implementation. Concerns around malpractice and the reluctance of doctors to rely on unfamiliar technology for critical decisions are among the key considerations.
A study analyzing Census Bureau data from January indicates that 6.1% of healthcare companies are planning to integrate AI within the next six months, reflecting a growing trend towards AI adoption.
AI systems, like any medical intervention, carry inherent risks to patients, including the potential for unforeseen consequences. Wachter recounts an incident where an AI-generated letter for a prescription inadvertently suggested an unconventional treatment due to misinterpretation by the system.
Furthermore, there are concerns about AI exacerbating existing disparities in healthcare, particularly in perpetuating discrimination against minority groups. Studies have shown instances where AI algorithms, when trained on biased data, can inadvertently reinforce discriminatory practices.
Addressing these challenges will require ongoing vigilance and oversight, emphasizing the importance of transparent algorithms monitored by human experts, regulators, and independent researchers. Continuous advancements in AI necessitate adaptive regulatory frameworks and proactive measures to mitigate risks and ensure accountability.
As the landscape of AI in healthcare evolves, the imperative for innovative monitoring systems and unforeseen advancements underscore the need for proactive legislative measures to anticipate and address emerging issues effectively.
Darius Tahir: @dariustahir, @kff.org
