The Consumer Technology Association (CTA®) has released new, industry-developed technology standards addressing the systemic challenges with bias in artificial intelligence (AI) used in health care. The consensus-based standards help AI product developers and risk managers identify the types, sources, and management practices of bias for AI tools.
CTA’s ANSI-accredited standard identifies the types of bias, where those biases can occur throughout the data lifecycle, and good data management strategies to minimize bias. Recognizing the potential pitfalls, the standard supports efforts by AI product developers and risk managers to minimize and effectively manage bias within AI systems.
"Incorporating AI into health care has incredible potential, but our industry has a responsibility to prioritize fairness, transparency, and well-being,” said Kerri Haresign, Senior Director of Technology and Standards at CTA. “By fostering unbiased AI development, we can maximize the benefits of this technology while upholding the highest standards of ethics and inclusivity. We are meeting a critical opportunity to set the conditions for sensible regulations that make the U.S. a global leader in AI policy and innovation."
The Artificial Intelligence in Health Care: Practices for Identifying and Managing Bias (CTA-2116) standard:
Categorizes different types of bias that can permeate AI systems, pinpointing the stages within the data lifecycle where these biases show up.
Recommends “Good Data Management Practices” to minimize bias and ensure that AI technologies are developed and implemented with utmost fairness and accuracy.
Supports AI product developers, risk managers, regulators, and health care practitioners who share the common goal of advancing health care through innovation, giving them tools to safeguard against potential bias.
With over 1000 active participants, CTA’s standards program creates the blueprints behind innovation in virtually every technology industry vertical. AI standards support innovation, and CTA has made early contributions by releasing publications regarding Definitions and Characteristics of Artificial Intelligence (CTA-2089-A), Best Practices and Recommendations for Bias Management (CTA-2116) and The Use of Artificial Intelligence in Health Care: Managing, Characterizing, and Safeguarding Data (CTA-2107-A).
CTA also recently released a National AI Policy Framework, which presents a roadmap for legislative action that allows the U.S. to lead responsibly and transparently on AI. The framework was accompanied by research showing consumers agree AI regulation is needed and feel that the U.S. federal government should lead on AI regulation by adopting legislation that balances consumer protection and the flexibility needed to unlock innovation.
About Consumer Technology Association:
As North America’s largest technology trade association, CTA® is the tech sector. Our members are the world’s leading innovators – from startups to global brands – helping support more than 18 million American jobs. CTA owns and produces CES® – the most powerful tech event in the world. Find us at CTA.tech. Follow us at @CTAtech.