Radical Reads

FDA’s Action Plan for Regulating AI/ML

By Sanjana Basu, Investor


Source: FDA Artificial Intelligence/Machine Learning (AI/ML)-Based Software as a Medical Device (SaMD) Action Plan, January 2021

The opportunity for artificial intelligence to transform healthcare was put in sharp relief in 2020, as AI research and applications were designed to help diagnose coronavirus infections, improve heart transplant outcomes, and detect arthritis before it develops. Newly announced regulatory guidelines aim to further reduce the obstacles to implementing AI technologies in daily clinical practice. Last week, the FDA issued its Artificial Intelligence and Machine Learning (AI/ML) Software as a Medical Device Action (SaMD) plan that outlines best practices for the application of AI and ML in healthcare.

We have discussed how AI will play a major role in nearly every aspect of healthcare, from diagnosis to treatment (see our framework on the two waves of digital health innovation). Within the second wave of innovation, we believe that prescription digital therapeutics and other AI/ML based diagnostic technologies will become mainstream. Defined as “software intended to be used for one or more medical purposes that perform these purposes without being part of a hardware medical device,” these technologies will lay the foundation for evidence-based medicine ranging from stroke detection to treatments of substance abuse disorders.

One of the key takeaways from the FDA discussion paper is the predetermined change control plan which expects transparency and real world performance monitoring from manufacturers, enabling the FDA to ensure safety and effectiveness across the product lifecycle. Given the adaptive nature of AI, these regulations may create a market for an additional set of technologies capable of continuously monitoring the algorithms and providing explainability about how the algorithms are making decisions.

In addition to how AI technologies are deployed within SaMDs, the FDA discussion paper touches on the need to hold the healthcare industry to a high standard on issues of bias and security. We are encouraged by the framework outlined by the FDA and welcome conversations about AI that move  healthcare forward while building patient trust.

AI News This Week

  • Canada CIFAR AI Chairs program surges past 100  (CIFAR)

    Canada welcomes 29 international researchers to its world-class AI research and training community to advance research in a wide range of areas, including machine learning for health and responsible AI. The recipients include Radical Scientific Advisor David Duvenaud. The Canada CIFAR AI Chairs program, a cornerstone of the Pan-Canadian AI Strategy (co-written by Radical and Vector Institute co-founders Tomi Poutanen and Jordan Jacobs), aims to attract outstanding researchers to Canada by providing them with long-term, dedicated funding to pursue innovative ideas. Disclosure: Jordan Jacobs is a Director of CIFAR.

  • AI Reflections in 2020  (Nature Machine Intelligence)

    A look at research trends concerning AI in society from the previous year and analysis of what some of the top academics in the field expect for 2021. The anticipated trends include growing demand for algorithmic transparency, investments in addressing bias in AI systems, and steps to increase governance of technology.

  • AI could make healthcare fairer – by helping us believe what patients say  (MIT Technology Review)

    A new paper proposes a way to develop medical algorithms that might reverse existing inequalities in healthcare delivery. When algorithms are trained only on expert knowledge, they are vulnerable to perpetuating existing biases. An experiment correlating patient pain with detailed radiology reports indicates that the standard way of measuring pain is flawed, revealing bias in the treatment of Black patients.

  • The New Architecture of Tech Innovation Looks Very Much Like This   (Forbes)

    Collaboration and partnerships can be the primary driver of innovation. ATB Ventures, an arm of ATB Financial, has partnered with Radical Ventures to advance ATB’s innovation efforts in applied AI and machine learning. The research coming out of Alberta’s AI ecosystem may play an important role in shaping our future economy.  ATB’s investment in Radical represents an important connection that builds on the recent growth and success of Alberta’s tech and venture capital ecosystem.

  • Why Amazon is interested in outside receipts   (Bank Administration Institute)

    Amazon recently launched a program that pays customers to upload their receipts from outside retailers, putting the tech giant in direct competition with banks. In this article, Sensibill CEO, Corey Gross, discusses how financial institutions can take proactive steps to protect their relationships, encourage loyalty and maintain market share. Sensibill is a Radical portfolio company.

Radical Reads is edited by Leah Morris (Senior Director, Velocity Program, Radical Ventures).