Creating a Strategy
Policy and strategy in the US
In an executive order titled “Accelerating America’s Leadership in Artificial Intelligence,” the U.S. government laid out its overarching strategy for AI policy with a focus on promoting AI development and innovation.¹ Five initiatives were described, including the creation of standards to guide research and development, bolstering the availability of resources necessary for AI growth, and developing an AI workforce. In line with this policy direction, a draft rule was released for comment in January 2020 that provides more specifics on the direction of government oversight for AI. The "Guidance for Regulation of Artificial Intelligence Applications" draft rule lists key considerations for other government agencies when creating AI regulation.² Among the top considerations are the need to promote public trust in AI and to minimize or eliminate regulation where possible.
Policy and strategy in the UK
In the UK, the National Health Service (NHS) developed a comprehensive document, "Artificial Intelligence: How to get it right (Putting policy into practice for safe data-driven innovation in health and care),” highlighting elements such as the governance of AI, ethical considerations, and adoption and spread.³ They preface this report with the need to take action in five key areas: leadership and society, skills and talent, access to data, supporting adoption, and international engagement. Of these, fostering a collaborative relationship among industry, academia, and government is highlighted to be absolutely critical to pave the way for adoption of AI in healthcare settings.
Market landscape and gaps in effective guideline implementation
The NHS also points to results from a 2019 survey, which looked at the proportion of products likely or very likely to be ready for at-scale deployment in 1, 3, or 5 years, which revealed an overall low market-readiness sentiment by AI developers. Understanding this, it is all the more important for stakeholders to work together to shape the future of AI, to ensure that datasets used in developing AI models are "FAIR (findable, accessible, interoperable, and reusable) and used appropriately."³ The authors pointed out that existing governance frameworks for AI technologies may be limiting innovation. Developers, when surveyed, stated that they were unaware that guidelines existed and about 50% did not have plans to certify their product as a medical device. Whatever the reason, these sentiments show that greater clarity is needed; otherwise, we risk having innovations being developed without proper consideration of issues such as bias and ethics, and ultimately, compromising patient and population safety.
Policy and strategy in the EU
The NHS AI report goes into impressive detail, which go beyond the scope of this series, but to give you a few examples, it includes topics such as the 'black box' nature of AI as well as case studies from the UK health system. As for the European Union, the European Commission put forth three pillars as part of their approach to AI; these include increasing investments to encourage public and private uptake, supporting education and training initiatives to develop and retain AI talent, and building an appropriate ethical and legal framework for AI.⁴ They published a white paper on artificial intelligence in February 2020. Compared to that of the FDA, the European report is broader in nature, encompassing climate change, agriculture, security, all in addition to healthcare.⁵ An emphasis is placed on people's trust of these technologies and how trustworthiness is a prerequisite for their uptake, and so, they stress that risks to fundamental rights (e.g. personal data and privacy protection, safety) must be mitigated. A common theme that has emerged among the various regulatory bodies is the need to develop a framework that is agile. They recognize that we live in a time in which technologies are rapidly improving and evolving, and so, the regulatory structure must also support that. As this structure continues to be developed, it must go beyond the implications on the individual, and focus on those on the society.
Key stakeholders and organizations
As we begin to see the implementation of artificial intelligence and machine learning within our healthcare systems, subject experts and reputable organizations have simultaneously been working on public policies and position statements. We wanted to share some highlights from these documents published by groups such as the American Medical Association (AMA), the American Society of Health System Pharmacist (ASHP), and Healthcare Information and Management Systems Society (HIMSS).
AMA's vision involves the integration of practicing physicians' perspectives into the development, design, validation, and implementation of AI in healthcare, and in doing so, keeping the user at the core of every thought process and solution. The development and design of any healthcare AI must be reproducible, be kept transparent, address biases, and protect patients, both in terms of health information and their well-being.⁶
ASHP is one of the only organizations to specifically address the role of the pharmacist in AI. In its Statement on the Use of Artificial Intelligence in Pharmacy, ASHP states that pharmacists should serve as key stakeholders in determining which aspects of medication use and management should be handled by AI and to what degree.⁷ AI has the potential to advance health and processes related to medication use, but it must be developed and deployed thoughtfully and strategically to ensure both safety and efficacy. Key questions that must be addressed include what medication-related tasks are appropriate for AI, how AI models should be evaluated, and what the most appropriate AI model and learning approach are for each use case. The statement maintains that pharmacists should play an integral role in the design, development, deployment, and continuous monitoring of AI systems, specifically as it relates to clinical applications, pharmacy operations, and pharmacy informatics.
Acknowledging the important current and future role of AI in diagnostics, treatment, care delivery, and care access, HIMSS maintains that AI cannot achieve its capacity for transforming healthcare without effective policy to guide implementation. It strives to promote the creation of effective policy that outlines where to strategically focus efforts that would enable proper AI integration.
Aside from the various medical associations, we also have grassroots organizations looking to bring artificial intelligence to more people across the globe, no matter their educational background. Fast.AI, as an example, aims to educate people on deep learning and simultaneously, make deep learning more accessible and easier to use.⁸ OpenAI is an AI research and deployment company whose mission is to ensure that artificial general intelligence benefits all of humanity.⁹ They published a charter document outlining their principles, which include committing to research required to make AGI safe, pushing for broader adoption of AI research, and collaborating with research and policy institutions.¹⁰ The Allen Institute for AI and Partnership on AI are two other examples, but more specific to healthcare, we have the Alliance for Artificial Intelligence in Healthcare (AAIH) and PathAI (PATH).¹¹⁻¹⁴. The AAIH is a global advocacy organization made up of representatives from the technology, pharmaceutical, and research sectors, who are working to establish standards for the use of AI in medicine and the overall healthcare system, while PATH is one that looks to explore the integration of automation, robotics, and AI from a healthcare perspective. What all these pioneer groups have in common is a mandate to support the use and development of AI in a safe and effective manner. To that end, you will find a number of quality publications, blog posts, and multimedia content on their websites that are intended to educate the community and to drive their mission.
Accelerating America’s Leadership in Artificial Intelligence. Published February 11, 2019. https://www.whitehouse.gov/articles/accelerating-americas-leadership-in-artificial-intelligence/
https://www.federalregister.gov/documents/2020/01/13/2020-00261/request-for-comments-on-a-draft-memorandum-to-the-heads-of-executive-departments-and-agencies
Joshi I, Morley J. Artificial Intelligence: How to get it right. Putting policy into practice for safe data-driven innovation in health and care. Published online 2019. https://www.nhsx.nhs.uk/media/documents/NHSX_AI_report.pdf
https://ec.europa.eu/digital-single-market/en/artificial-intelligence
European Commission. February 2020. White paper on artificial intelligence--a European approach to excellence and trust. Published online 2020. https://ec.europa.eu/info/sites/info/files/commission-white-paper-artificial-intelligence-feb2020_en.pdf
American Medical Association. Augmented Intelligence in Health Care policy report. 2018. https://www.ama-assn.org/system/files/2019-01/augmented-intelligence-policy-report.pdf
American Society of Health-System Pharmacists. Statement on the Use of Artificial Intelligence in Pharmacy. 2020. https://www.ashp.org/-/media/assets/policy-guidelines/docs/statements/artificial-intelligence-in-pharmacy.ashx?la=en&hash=CD600F3261674788293EAB37074F1352E85CFECA.
https://www.fast.ai/
https://openai.com/
https://openai.com/charter/
https://allenai.org/
https://www.partnershiponai.org/
https://www.theaaih.org/
https://pathhealth.com/