发表机构
Aegis Compliance and Strategies OÜ（Aegis合规与策略有限责任公司）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究揭示生成式AI下专家服务责任承诺面临三重天花板，证明可证明性依赖状态，并校准显示2026年保险除外条款将显著改变约束格局。
AI 中文摘要
当生成式人工智能将具有说服力的专家制品的边际成本推向零时，基于生产成本的胜任力信号崩溃，取而代之的是基于结果的或有责任承诺。此类承诺看似对更强大的人工智能具有稳健性：正失败率总会留下可定价的残差。我们表明，这种稳健性依赖于一个未言明的同质性假设，且责任承诺面临三重天花板，而非一重。人工智能错误可分解为特质性成分和共同成分；能力增长更快地消除特质性成分（Kim等人，2025），因此残差逐渐变得共同化——而共同错误恰恰是来自同一基础模型的验证者无法观察到的：评委评分与评委-生成器模型相似性之间的平均相关系数为r=0.84（Goel等人，2025）。因此，可证明性是状态依赖的，theta_eff = theta_0(1 - xi*kappa)。分离要求必须提交的承诺低于可提交的承诺，v/theta_eff <= min{Lbar, mv}，得出生存条件xi*kappabar <= 1 - 1/(M*theta_0)，其中M = min{Lbar/v, m}，并在票券规模v* = Lbar/m处发生制度切换：小规模业务受罚则原则约束，大规模业务受资本约束。针对四个司法辖区的六种业务类型进行校准后，自2026年1月生效的生成式人工智能保险除外条款将二十四个单元中的八个从罚则约束转变为偿付能力约束，并将xi=0.4下存活的单元从十二个削减至四个。保险撤出是大陆法系事件；普通法系单元已处于边界。
英文摘要
When generative AI drives the marginal cost of a persuasive expert artefact toward zero, production-cost signals of competence collapse and outcome-contingent liability commitments take their place. Such commitments look robust to better AI: a positive failure rate always leaves a residual to price. We show that this robustness rests on an unstated homogeneity assumption, and that a liability commitment faces three ceilings, not one. AI error decomposes into idiosyncratic and common components; capability growth eliminates the idiosyncratic part faster (Kim et al., 2025), so the residual becomes progressively common - and common error is precisely what a verifier drawn from the same foundation model cannot observe: judge scores correlate with judge-generator model similarity at an average r=0.84 (Goel et al., 2025). Provability is therefore state-dependent, theta_eff = theta_0(1 - xi*kappa). Separation requires the commitment that must be posted to fall below what can be posted, v/theta_eff <= min{Lbar, mv}, yielding the survival condition xi*kappabar <= 1 - 1/(M*theta_0) with M = min{Lbar/v, m} and a regime switch at the ticket size v* = Lbar/m: small engagements are constrained by the penalty doctrine, large ones by capital. Calibrated to six engagement types across four jurisdictions, the generative-AI insurance exclusions effective January 2026 shift eight of twenty-four cells from doctrine-bound to solvency-bound and cut the cells surviving xi=0.4 from twelve to four. The insurance withdrawal is a civil-law event; the common-law cells were already at the boundary.
Comments23 pages, 2 figures. Part of a working-paper series on liability signalling under improving AI (arXiv:2607.26327, arXiv:2608.04276, arXiv:2608.05969). Replication package: doi:10.6084/m9.figshare.33188046; companion verification package: doi:10.6084/m9.figshare.33190722
