AccountabilityTransparencyTrustworthinesscs.AI
Can AI be Accountable?
Andrew L. Kun
October 30, 2025
arXiv.org
1 citations
Key Takeaways
-
Accountability requires three components: For AI to be truly accountable, affected parties must be able to request information about the AI's actions, discuss that information with developers or operators, and impose consequences when needed—yet many current AI systems fail to meet these basic requirements.
-
Current AI lacks transparency and responsiveness: The paper identifies a critical gap where today's AI cannot be meaningfully questioned or engaged in dialogue about its decisions, making it impossible for users, consumers, or regulators to hold it responsible for its impacts.
-
Accountability is essential for public trust: As AI becomes more powerful and affects more decisions in people's lives, establishing accountability mechanisms is described as "imperative" for ensuring AI actually serves the needs of consumers, voters, and decision-makers rather than operating as a black box.
-
Solutions exist but require deliberate implementation: The paper suggests that approaches to improve AI accountability are available, but achieving a world where all AI is accountable to those affected by it requires intentional effort and systemic change rather than happening automatically.
Abstract
The AI we use is powerful, and its power is increasing rapidly. If this powerful AI is to serve the needs of consumers, voters, and decision makers, then it is imperative that the AI is accountable. In general, an agent is accountable to a forum if the forum can request information from the agent about its actions, if the forum and the agent can discuss this information, and if the forum can sanction the agent. Unfortunately, in too many cases today's AI is not accountable -- we cannot question it, enter into a discussion with it, let alone sanction it. In this chapter we relate the general definition of accountability to AI, we illustrate what it means for AI to be accountable and unaccountable, and we explore approaches that can improve our chances of living in a world where all AI is accountable to those who are affected by it.