Minimum Viable Ethics: From Institutionalizing Industry AI Governance to Product Impact

AI ethics roles exist but lack real power: The study found that despite companies creating dedicated AI ethics positions, these professionals have "little traditional power" and must work around their lack of authority over actual product decisions, relying instead on informal influence and negotiation tactics.

Ethics gets reduced to compliance and quality control: The research reveals that AI ethics work in industry has narrowed to a "minimum viable ethics" focused primarily on regulatory compliance and product quality assurance, rather than addressing broader normative questions about what is right or fair—limiting the scope of ethical considerations.

Ethics professionals use regulatory fear as leverage: When negotiating with product teams, AI ethics professionals strategically invoke "narratives of regulatory response" to advance their work, suggesting that the threat of future regulation is often more persuasive than ethical arguments alone in driving corporate action.

Current industry approach may be insufficient: The paper suggests that relying on internal ethics professionals without formal authority has created a system where meaningful product changes remain unclear, and that "potential future regulation may help bridge this gap"—implying voluntary corporate ethics efforts alone may not be adequate.

Ethics work requires creative workarounds: The study found that AI ethics professionals are "highly agile and opportunistic," developing informal strategies and "standardized and reusable processes and tools" to advance ethics work within corporate constraints, rather than having formal mechanisms to enforce ethical considerations.