The silence around AI accountability is not an accident. It’s a choice. Governments and corporations have spent years drafting frameworks, convening ethics boards, and issuing white papers on responsible development. Yet one question remains unanswered in every document, every summit, every press release: who is legally liable when an AI system causes harm? The absence of an answer is not oversight. It’s a deliberate buffer.
Consider the pattern. When a car fails and injures someone, the manufacturer can be sued. When a drug has undisclosed side effects, the pharmaceutical company faces litigation. But when an AI system denies a loan, misdiagnoses a patient, or amplifies disinformation, the chain of responsibility dissolves. Developers point to users. Users point to opaque algorithms. Regulators shrug and say the tech moves too fast for law to keep up. The result is a vacuum where accountability should be. No fines. No penalties. No precedent. Just a growing pile of case studies labeled “lessons learned.”
Looking at the Record
The European Union’s AI Act, hailed as a landmark in governance, categorizes systems by risk and imposes strict rules on high-risk applications. But it stops short of defining personal or corporate liability for failures. The U.S. executive order on AI safety from October 2023 mandates reporting requirements for powerful models. Yet it sidesteps the question of who pays—literally or figuratively—when those models err. Even voluntary industry commitments, like the ones signed at global AI summits, lean on vague promises of “responsibility” without binding mechanisms. The word “liability” rarely appears. When it does, it’s buried in footnotes or deferred to future committees.
The Consequences of Silence
This silence has consequences. Without legal stakes, there’s no incentive to prioritize safety over speed. Companies race to deploy, knowing the worst outcome is a public apology or a congressional hearing they can weather with PR. Meanwhile, individuals harmed by AI decisions—denied jobs, misidentified by facial recognition, targeted by biased content—have no clear path to recourse. The cost falls on them, not the creators. This isn’t hypothetical. Studies from 2022 showed over 60% of U.S. adults surveyed by Pew Research worried about AI-driven discrimination in hiring. Court cases are piling up, but rulings are inconsistent because the law itself is silent.
An Intentional Calculation
The implication is clear. By avoiding liability rules, policymakers and industry leaders are deciding that AI’s benefits outweigh its harms—or at least, that the harms can be externalized. It’s a calculation, not a gap. If accountability were a priority, we’d see draft bills naming specific penalties, not just guidelines. We’d see regulators demanding transparency on decision-making processes, not just on training data. Instead, the focus stays on innovation, on keeping the field “competitive.” Harm becomes collateral.
Note for the archive: silence on liability isn’t neutrality. It’s a policy stance. It says the burden of proof—and the burden of loss—belongs to the public, not the powerful. That’s not a glitch in the system. It’s the system working as designed.
