By: Scott Bonney, VP of Health IT Services at RELI Group
Federal agencies have engaged in a bevy of artificial intelligence (AI) initiatives, policies and guidance following the White House’s landmark AI executive order, but there’s been little in the way of a government-wide approach to the emerging technology as it gets integrated into nearly every part of operations.
In March, the Office of Management and Budget (OMB) became the first federal agency to issue comprehensive guidance, which established additional requirements and direction for AI governance, innovation and risk management for uses that impact the rights and safety of citizens.
Given the amount and nature of the data entrusted to government agencies, specifically federal health agencies, it’s imperative to implement a government-wide approach for the use of AI to ensure agency leaders are taking every possible precaution to secure citizen data and privacy.
As outlined in the White House’s December 2023 memo, AI will undoubtedly impact health agencies along with the care and services citizens receive, so building public trust and confidence in the technology is more important than ever as agencies seek to expand its use across federal health missions.
Endless Possibilities for AI in Healthcare
With wider adoption of AI, it’s estimated the technology could lead to a five to ten percent savings in U.S. healthcare spending, roughly $200 to $360 billion. Not only can AI benefit ongoing healthcare initiatives such as research, fraud prevention and data collection, it has tremendous potential to improve health outcomes and care for citizens.
Since government leaders are still in the very early stages of understanding and implementing AI, identifying use cases to demonstrate the benefits of the technology — and that it does not endanger the rights and safety of citizens — is essential for success.
For example, during the COVID-19 pandemic, the National Center for Immunization and Respiratory Diseases (NCIRD) at Centers for Disease Control and Prevention (CDC) developed a conversational chatbot to analyze questions submitted by the public, healthcare providers, partners or internal staff to issue agency-approved answers that provide the best support. More than 2,300 users who accessed the chat tool were able to gain information regarding where to receive a vaccine or booster, empowering those with the information needed to seek care.
Additionally, AI can improve medical diagnosis by improving accuracy and reducing time of interpretation. Manually analyzing data sets and results can take anywhere from days to months, but using an AI algorithm to analyze data and identify patterns can enable providers to proactively identify a health issue. With that extra time, providers can focus more on their plan of action to remediate an illness versus simply identifying it.
The benefits of AI in providing higher quality healthcare for citizens are endless, but to take full advantage, agency leaders must determine how to secure and maintain citizen trust to reap the full benefits.
Create a Foundation for Trust
OMB’s recent guidance emphasizes the continuous process of risk management through specific minimum practices for uses of AI that impact the public’s rights and safety. This especially applies to healthcare, as government leaders work to protect patient information and improve access to care.
The importance of healthcare cybersecurity is highlighted by the creation of the Health Sector Cybersecurity Coordinator Center (HC3). From 2018 to 2022, government research showed a 93% increase in large data breaches, with a 278% increase in large breaches reported involving ransomware. There is a clear need for risk management and security best practices when incorporating new technologies into existing infrastructure.
The first step to adopting AI safely is understanding where it fits into security cycles or mission processes. While federal agencies help citizens address a variety of needs, the healthcare field remains a top priority with unique concerns. As a result, unique AI models and systems for healthcare must be developed to ensure successful patient outcomes. To ensure the highest level of security, AI systems must be designed and implemented according to sector to provide the appropriate functions and security features.
Taking advantage of the benefits of automation also enables thorough compliance and risk management. Tools such as ServiceNow can help organizations implement governance, risk, and compliance (GRC) tracking and awareness, contributing to the overall goal of AI safety.
In addition, agencies must always utilize common cyber defense best practices and disclosures, including VPNs, using complex passwords, communicating regularly with cybersecurity teams and patient awareness of AI systems in use. Incorporating risk management and transparency in agency processes is necessary to build citizen trust, especially when using technologies such as AI that pose many questions for citizens.
Leaders should also consider the decision processes affected by AI use. AI should never inform with built–in bias or discrimination, even if unintentional, in order to ensure fairness and transparency. Responsible AI usage often necessitates human oversight to verify the results produced by AI tools. In healthcare, for instance, CMS recognizes that human oversight is crucial to ensure AI applications promote equitable access to care in prior authorizations and medical necessity determinations.
Finally, AI systems are only as accurate and helpful as the data provided, which highlights why secure data is needed for successful, accurate outcomes. If an AI model is being used to predict a health outcome for instance, and it is provided with inaccurate, tampered data, then the resulting outcome will produce poor results. Utilizing secure, accurate and non-biased data is crucial to AI’s success.
The use of emerging technologies such as AI to improve health outcomes is exciting, even enticing, but it is government’s responsibility to alleviate any public concerns and apprehension. As long as agencies remain transparent with citizens and communicate their best practices, AI has the potential to transform patient care and the healthcare sector.
