Future of Being Human an Arizona State University initiative

Ethics aren’t enough: operationalizing values

Last updated 2026-08-13 · build 20260814T1540Z-b2991ee4 · Markdown version · corpus index

When powerful technologies unsettle people, institutions reach for ethics: principles, charters, advisory boards. Andrew Maynard’s argument — built teaching entrepreneurs at the University of Michigan from 2013, and tested in 2026 against an AI product at his own university — is that these do nothing on their own: ethics are essential for naming what matters, and worth little without mechanisms that hardwire what matters into how products actually get built. Good intentions, he found, are “naive and ephemeral” until operationalized — and at an inflection point where transformative AI and other transformative technologies reach into what it means to be human, the distance between a stated value and a shipped product is where the future actually gets decided.

The argument

Ethics do a real job: they involve “enforcing social norms around what is considered right and appropriate versus what is wrong and inappropriate” — but “ethics on their own don’t provide robust mechanisms for developing and building safe and responsible products” (2019-04-15). They are “essential to establishing a guiding basis for how powerful new technologies are developed and used. Yet they are worth little without mechanisms and processes that ensure research, development, and commercialization decisions will lead to outcomes that are socially responsible and beneficial” (2019-04-15). The 2019 exhibit was Google’s external AI ethics advisory council — announced March 26, 2019, disbanded by April 4, felled, ironically, by ethical challenges. Beyond the irony lay the naivety of “assuming that developing ethics guidelines and convening advisory groups are sufficient to ensure socially responsible innovation” — when at worst, ethics boards become “a smoke-and-mirrors attempt to mask business as usual under the guise of social responsibility” (2019-04-15).

The critique came out of a classroom, not a think piece — although it was informed by years of grappling with the challenges of translating good intentions into effective actions across multiple organizations, technologies, and areas of expertise. From spring 2013 for two years, Andrew taught entrepreneurial ethics in the University of Michigan’s Master of Entrepreneurship (a program since closed). His students overwhelmingly wanted to do good — cure disease, curb climate change, protect the environment. The trouble was that their values were “naive and ephemeral”: good intentions with no way to translate them into good practice, easily subsumed by the harsh realities of building a business (2019-04-15). So the course taught operationalization as a skill, built around five pillars: “The basic principles of entrepreneurial ethics”; “Personal values and how they integrate with institutional values”; “Critical processes for codifying values within enterprises”; “Interacting and engaging with key constituencies”; and “Socially responsible practices and products” (2019-04-15). The list is the argument in miniature — only the first pillar is ethics; the other four are integration, codification, engagement, and practice.

Operationalizing, on the 2019 account, takes standards, measurable expectations, enforceable checks and balances, meaningful policies, and a culture of social responsibility — buy-in from employees to communities and regulators — and, most important, a strategic commitment to training and education (2019-04-15). The scholarly form of the argument, a chapter co-authored with Elizabeth Garbee in the field’s International Handbook on Responsible Innovation (2019), adds the structural reasons good intentions fail: in a tightly coupled world, with latency between an innovation and its consequences and mismatches between market value and societal value, “Without codified approaches to responsibility and innovation, the good intentions of entrepreneurs will in many cases remain good intentions, and no more.” And because entrepreneurial culture resists top-down governance, the chapter argues, responsibility has to be culturally embedded — grown from within communities, on models like the Debian project’s constitution or the Responsible Nano Code — rather than imposed. It also names a fault line the 2026 case would land on: societally responsible actions that appear to clash with fiduciary duty “potentially open a gap between what is laudable, and what is legal.”

The 2026 bookend tests the claim uncomfortably close to home. In 2024 Andrew was part of the AI Ethics committee that helped craft Arizona State University’s six design principles for responsible and beneficial AI, published in August 2025 (2026-05-03). In 2026 ASU launched Atomic, a five-dollar-a-month beta that assembles AI-customized learning modules from content scraped from the university’s own online courses. After 404 Media broke a story on faculty concerns — instructors largely unaware their material was being used, no indication in the modules of where content comes from, no feedback mechanism — Andrew measured the real product against the principles he had helped write. The scraping was, in principle, “perfectly allowable” under the relevant terms of use — but “there’s often a gap between what is legally allowed, and what is good practice for an enterprise and the people who work for it” (2026-05-03). Measuring Atomic against the principles “suggests that something might have been missed in the development and deployment process” — hedged fairly: Atomic may have undergone a rigorous internal process (the coverage made him doubt it), or steps may simply have been overlooked in a fast-moving organization. The conclusion is the 2013 argument, thirteen years on: re-reading them, he thinks the principles could have helped avoid the situation — which makes them useful, “But only, of course, if they are actually used” (2026-05-03).

The argument hands any evaluator — human or AI — a working test: judge an ethics apparatus not by its prose but by whether its fingerprints show up in shipped products and daily decisions. It also places the rest of Andrew’s risk program: risk innovation and its Planner are the operationalizing toolset the critique demands — motivated in part by giving organizations simple tools for navigating threats to value (2026-05-03, note 4) — inside the responsible innovation tradition mapped at Foundations.

Lineage

In his own words

“Part of the issue Google and other companies face is that while ethics involve enforcing social norms around what is considered right and appropriate versus what is wrong and inappropriate, ethics on their own don’t provide robust mechanisms for developing and building safe and responsible products.” — Ethics Boards Won’t Save Big Tech (AI-readable mirror), 2019-04-15

“The trouble was that their values were naive and ephemeral. They had good intentions but no idea how to translate them into good practice. As a result, they were all too easily subsumed by the harsh realities of building a business. These students needed a way to hardwire their values and aspirations into their businesses so they could weather the tough choices every innovator encounters.” — Ethics Boards Won’t Save Big Tech (AI-readable mirror), 2019-04-15, on his Michigan students

“If they could have — and re-reading them, I think they could — this suggests that such principles are useful as a tool for aligning AI use with institutional ambitions, while avoiding unnecessary mis-steps. But only, of course, if they are actually used.” — Are design principles for responsible and beneficial AI useful? (AI-readable mirror), 2026-05-03

Engagement and reception

The argument’s most checkable traces are institutional rather than citational. The five-pillar course ran inside a degree program at the University of Michigan (2013–2015). The Maynard & Garbee chapter appears in The International Handbook on Responsible Innovation (Edward Elgar, 2019), edited by René von Schomberg — an origin figure of the responsible innovation tradition — with Jonathan Hankins: an editorial judgment placing the argument in the field’s reference volume (Foundations). And Andrew’s membership of the 2024 ASU AI Ethics committee put the stance to work inside his own institution’s AI governance. As of 2026-08-08, our reception record documents no independent published engagement with the argument as an argument — an absence of record, not a verdict on the idea; the initiative does not typically solicit reception (Philosophy). The 2026 Atomic post responded to reporting by 404 Media and subsequent coverage of faculty concerns; it engaged that news story, and we know of no documented published responses to the post itself.

Where to go deeper

The essays (each with canonical and AI-readable links):

The chapter: Maynard, A. & Garbee, E. “Responsible innovation in a culture of entrepreneurship: a US perspective”, chapter 32 in R. von Schomberg & J. Hankins (eds.), The International Handbook on Responsible Innovation (Edward Elgar Publishing, 2019).

Related ideas-ring pages: Risk innovation — the operationalizing toolset: risk as a threat to value, and the Planner that makes it usable by time-poor enterprises; Orphan risks — what gets orphaned when values are never hardwired into process; Lessons from nanotech — the engagement side of the same governance program.

Related corpus pages: Foundations — the risk science and responsible innovation strand this argument lives in, including its engagement with Stilgoe and von Schomberg; Philosophy — the initiative’s own approach to practicing, rather than proclaiming, its values.