In recent years, and even in recent months, artificial intelligence tools have rapidly expanded in their capabilities and use, especially in sectors like healthcare. Many have praised this technology for its ability to solve many of the biggest problems that the healthcare industry faces today, such as a severe staffing shortage that leaves workers under-resourced and overworked; however, others have called attention to the potential ethical implications that the use of this technology could have and the effects it could have on patients.
There are many applications for AI technology in the healthcare sector, ranging from simple to complex tasks. Some involve basic automation, such as sorting data or generating after-visit summaries and patient notes, while others are more complex, such as analyzing diagnostic imaging to provide a preliminary diagnosis or supporting researchers in drug development.
The goal of these applications is to accelerate and streamline workflows for healthcare professionals, allowing them to focus more on what matters most: providing patient care and saving lives.
However, this does not mean we should adopt AI unquestioningly without considering its potential consequences. Like any innovation, AI has its shortcomings, and it is important to engage with and mitigate these challenges — particularly in a field as sensitive as healthcare, where the stakes of many of the decisions AI is increasingly involved in are literally life and death.
How AI governance can help artificial intelligence be used more responsibly in healthcare
You might wonder why AI governance is necessary. Well, artificial intelligence is such a novel, emerging technology that government regulation simply cannot keep up with the pace of innovation.
By the time legislators agree on a new law, many of those provisions can be out of date, which is why internal governance has become so important. By allowing people who actually work with and understand AI to dictate the ways in which this technology can be ethically and responsibly deployed, we can move toward a future where AI is used for the benefit of the greater good.
Recently, we have seen a wave of institutions formalizing responsible AI frameworks for clinical use. These institutions range from academic institutions to coalitions of medical companies and institutional review boards (IRBs). However, it is important to note that these frameworks are not the be-all-end-all of AI governance, but instead a starting point to build upon as AI continues to develop and expand its reach.
One key example of AI governance in action comes from a collaboration between the Joint Commission and the Coalition for Health AI (CHAI). These organizations came together to develop the Guidance on the Responsible Use of Artificial Intelligence in Healthcare (RUAIH), the first formal framework for medical AI issued by a US accrediting body.
The framework outlines seven core areas necessary for responsible AI deployment in healthcare, which include:
- AI policies and governance structures
- Patient privacy and transparency
- Data security and data use protections
- Ongoing quality monitoring
- Voluntary reporting of AI safety-related events
- Risk and bias assessment
- Education and training
Another framework has been proposed by the Council for Responsible Use of AI in Clinical Trials, led by clinical research IRB Advarra, to “develop practical, consensus-driven guidance for responsible AI adoption in clinical research.” The goal of this framework is to effectively manage AI systems by evaluating the level of autonomy and degree of patient impact, with decisions being made based on these factors that work in tandem. For example, AI in a use case with high patient impact should be given low levels of autonomy.
Finding responsible and ethical ways to deploy AI in healthcare
One of the biggest concerns that critics have levied against the use of AI in healthcare is data privacy and security. As AI becomes increasingly prevalent in these settings, data oversight is becoming more central to these digital transformation strategies. Users of AI systems in healthcare settings must ensure that they pay attention not only to the security of systems to protect data but also to the quality of the data to prevent problems like algorithmic bias and information silos.
By embracing responsible AI governance policies like the ones that have recently been proposed, we can work towards a future where healthcare AI can be deployed in a scalable and ethical way. AI can lead to genuinely better patient outcomes while making workflows more efficient and streamlined for healthcare workers, but the real dangers of this technology must be acknowledged and prepared for by the healthcare system.
