PII Comment from Critchfield, Margot

Dear Esteemed Commissioners, I am an Episcopal priest with a passion for insuring that future generations are not denied a full, robust, and truthful accounting of our country’s founding and history, as I was. I had no idea until I was in my sixties that my understanding of the world, my knowledge of US history, and my perspective on all of life was so small and narrow, so limited, by the white lens through which all of it was filtered. They say you don’t know what you don’t know; it is equally true that once you see it, you can’t unsee it. And this is why I’m taking the time to submit this comment in strong opposition to the FTC’s proposed policy statement on AI accuracy, Matter No. P264200 / Docket FTC-2026-0859. This proposed statement does not protect consumers from AI harm, in fact it perpetuates it. It protects AI companies from accountability for discriminatory and structurally harmful outputs by reframing equity-centered correction and anti-discrimination safeguards as potential consumer deception. But any AI that hasn’t had the benefit of a multitude of diverse knowledge bases and diverse perspectives is by definition intellectually anemic. Remember that these systems are not neutral. They are trained primarily if not exclusively on white-centered data that is shaped by historical inequality, colonial archives, racial hierarchy, gender hierarchy, ableist assumptions, language dominance, and institutional exclusion. When those systems are educating us, that education is by definition intellectually anemic. And when used to inform consequential decisions, the harms are concrete. They affect jobs, housing, credit, healthcare, education, public benefits, legal systems, immigration, and personal safety. Notably, the FTC’s prior interagency position recognized that indeed automated systems can perpetuate unlawful bias, unlawful discrimination, and produce harmful outcomes. The Commission must not adopt a policy now that treats correction of those harms as suspicious. The proposed policy statement’s treatment of Colorado’s AI accountability law is dangerous. State laws that require companies to address algorithmic discrimination are not ideological coercion. They are consumer protection. They are civil-rights protection. They are public accountability. Section 5 of the FTC Act must not be used to preempt state-level AI accountability and anti-discrimination protections. State protections against algorithmic discrimination do not conflict with consumer protection. They are a necessary form of consumer protection. The proposed statement does not center the communities most harmed by uncorrected AI bias, including people of the global majority, disabled people, immigrants, LGBTQIA+ communities, women, workers, tenants, patients, students, and people targeted by surveillance and criminal legal systems. I urge the Commission to withdraw this proposed policy statement. I urge the Commission to authorize independent research on the documented harms of algorithmic discrimination. I urge the Commission to develop any future AI accuracy guidance through a process that centers the communities most harmed by AI systems, not the companies most protected by them. Sincerely, Rev. Margot Critchfield Sandwich, MA