{
 "failure_mode": "ai-oversight-and-alignment-gap",
 "specific_names": [
  "Adversarial feedback provider incentives",
  "Algorithmic Collusion",
  "Algorithmic Collusion Risk",
  "Closed model opacity",
  "Compound RLHF failure",
  "Explainable RL immaturity",
  "Helpfulness harmlessness tension",
  "Human unintelligibility of algorithmic optimization",
  "Inevitable corruption of unsupervised coded automation",
  "Majority capture of the reward model",
  "Overspecific human guidance",
  "Unquantifiable risk of centralized automation",
  "absence of legacy dynamic toolset",
  "absent dynamic governance toolset",
  "accountability gap without central control",
  "adversarial-agent-blindspot",
  "autonomy-divergence",
  "black box opacity",
  "circular-dependency",
  "compliance-circularity",
  "constraint circumvention by capability growth",
  "escalating alignment tax",
  "federated accountability gap",
  "human oversight bias recursion",
  "human-in-the-loop cost drag",
  "illusory alignment",
  "implementation gap in bias audits",
  "incomplete accountability under human-in-the-loop",
  "internal-loop-recalibration",
  "missing-feedback-mechanisms",
  "monitoring obsolescence",
  "no-feedback-no-anticipation",
  "opacity undermining procedural fairness",
  "opaque high stakes decisioning",
  "preemptive anticipation limit",
  "reactive governance insufficiency",
  "reactive-not-proactive",
  "self-monitoring-collusion",
  "self-referential-monitoring-loop",
  "self-simulated-threat-model",
  "static-iot-linkage",
  "static-monitoring-services",
  "unspecified-safeguards-in-self-monitoring",
  "unstandardized-internal-oversight",
  "unverified post hoc explanations",
  "validation layer scope limit"
 ],
 "count": 47,
 "claims": [
  {
   "id": "kaal:claim:3808867-023",
   "url": "https://wulfkaal.github.io/claims/3808867-023",
   "claim": "The conveniences and benefits of centralized algorithmic automation carry risks to humanity that cannot be fully quantified, and decentralized systems can counteract those downsides and threats.",
   "specific_name": "Unquantifiable risk of centralized automation",
   "conditions": [],
   "source": "How Decentralized Systems Can Upgrade AI",
   "year": "2021",
   "quote": "At the same time, the conveniences and benefits that derive from centralized algorithmic automation bring with them risks to humanity that cannot be fully quantified. Decentralized systems can counteract the downsides and threats of algorithmic automation",
   "citation": "Wulf A. Kaal, How Decentralized Systems Can Upgrade AI (2021). SSRN: https://ssrn.com/abstract=3808867"
  },
  {
   "id": "kaal:claim:3808867-025",
   "url": "https://wulfkaal.github.io/claims/3808867-025",
   "claim": "The quantification and algorithmic optimization of human thought, feeling, and action can produce an optimization of humans that is too complex for humans to understand without the data driven algorithmic aids themselves.",
   "specific_name": "Human unintelligibility of algorithmic optimization",
   "conditions": [
    "sensor and wearable derived data on human interaction"
   ],
   "source": "How Decentralized Systems Can Upgrade AI",
   "year": "2021",
   "quote": "The quantification of human thought, feeling, and action and their algorithmic optimization can create optimization of humans that is too complex for humans to understand without the data-driven algorithmic aids.",
   "citation": "Wulf A. Kaal, How Decentralized Systems Can Upgrade AI (2021). SSRN: https://ssrn.com/abstract=3808867"
  },
  {
   "id": "kaal:claim:4796714-002",
   "url": "https://wulfkaal.github.io/claims/4796714-002",
   "claim": "Conventional governance methods that are reactive or fixed to ex-post solutions are insufficient for governing technologies whose behavior changes continuously after deployment.",
   "specific_name": "reactive governance insufficiency",
   "conditions": [
    "applies to ex-post regulatory regimes",
    "holds for AI systems that continue to learn after release"
   ],
   "source": "AI Governance",
   "year": "2024",
   "quote": "These conventional methods, largely reactive or fixed to ex-post solutions, are proving insufficient for the dynamic nature of AI technologies.",
   "citation": "Wulf A. Kaal, AI Governance (2024). SSRN: https://ssrn.com/abstract=4796714"
  },
  {
   "id": "kaal:claim:4796714-003",
   "url": "https://wulfkaal.github.io/claims/4796714-003",
   "claim": "At the time of publication no legacy governance system exists that can supply the dynamic governance toolsets required to govern evolving AI models ex-ante and manage deployed solutions ex-post.",
   "specific_name": "absence of legacy dynamic toolset",
   "conditions": [
    "as of the paper's publication in 2024",
    "refers to legacy, non web3 governance infrastructure"
   ],
   "source": "AI Governance",
   "year": "2024",
   "quote": "No legacy systems exist at the time of publication that could provide such dynamic governance toolsets.",
   "citation": "Wulf A. Kaal, AI Governance (2024). SSRN: https://ssrn.com/abstract=4796714"
  },
  {
   "id": "kaal:claim:4796714-007",
   "url": "https://wulfkaal.github.io/claims/4796714-007",
   "claim": "The black box character of deep learning models is a governance failure and not merely a technical inconvenience: opacity obstructs debugging, obscures bias detection and mitigation, and prevents comprehension of how inputs become outputs.",
   "specific_name": "black box opacity",
   "conditions": [
    "most pronounced in deep learning models",
    "matters where decisions must be explained or audited"
   ],
   "source": "AI Governance",
   "year": "2024",
   "quote": "This opacity can obstruct the debugging process, obscure bias detection and mitigation, and hinder comprehension of AI decision-making.",
   "citation": "Wulf A. Kaal, AI Governance (2024). SSRN: https://ssrn.com/abstract=4796714"
  },
  {
   "id": "kaal:claim:4796714-017",
   "url": "https://wulfkaal.github.io/claims/4796714-017",
   "claim": "Using human judgment to uncover unconscious bias in AI can perpetuate the very biases it is meant to remove, because human reviewers carry their own implicit biases and may lack the expertise to identify bias in complex AI systems.",
   "specific_name": "human oversight bias recursion",
   "conditions": [
    "applies to human in the loop bias mitigation programs",
    "holds where reviewers lack technical expertise or bias awareness"
   ],
   "source": "AI Governance",
   "year": "2024",
   "quote": "While human judgment is integral to risk management and bias mitigation, it inherently carries its own biases. Relying on human judgment to uncover unconscious biases in AI may inadvertently perpetuate these biases rather than eliminate them.",
   "citation": "Wulf A. Kaal, AI Governance (2024). SSRN: https://ssrn.com/abstract=4796714"
  },
  {
   "id": "kaal:claim:4796714-023",
   "url": "https://wulfkaal.github.io/claims/4796714-023",
   "claim": "Purely preemptive regulation cannot succeed on its own, because it is not possible to anticipate every issue or bias an AI system will exhibit before it is operational and interacting with real world variables.",
   "specific_name": "preemptive anticipation limit",
   "conditions": [
    "applies to ex-ante controls imposed at the model and data levels",
    "concerns emergent behavior that appears only in deployment"
   ],
   "source": "AI Governance",
   "year": "2024",
   "quote": "It is challenging to anticipate all potential issues or biases that may arise with an AI system before it is fully operational and interacting with real-world variables. Regulations that insist on preemptive controls may fail to address unforeseen problems that only become evident after deployment.",
   "citation": "Wulf A. Kaal, AI Governance (2024). SSRN: https://ssrn.com/abstract=4796714"
  },
  {
   "id": "kaal:claim:4796714-027",
   "url": "https://wulfkaal.github.io/claims/4796714-027",
   "claim": "Post-deployment monitoring, the standard fallback when ex-ante rules prove inadequate, is typically woefully outdated by the time it is applied because the AI models continue to evolve.",
   "specific_name": "monitoring obsolescence",
   "conditions": [
    "applies to monitoring regimes applied after deployment",
    "holds for continuously learning models"
   ],
   "source": "AI Governance",
   "year": "2024",
   "quote": "Yet, post-deployment monitoring is typically woefully outdated as the AI models evolve.",
   "citation": "Wulf A. Kaal, AI Governance (2024). SSRN: https://ssrn.com/abstract=4796714"
  },
  {
   "id": "kaal:claim:4796714-030",
   "url": "https://wulfkaal.github.io/claims/4796714-030",
   "claim": "Transparency and accountability cannot be assured across all participants in a federated governance model because there is no centralized control, and the author declines to advocate centralized control as the remedy; the consequences are biased or unfair AI systems, inadequate privacy protection, and unequal access to AI benefits.",
   "specific_name": "federated accountability gap",
   "conditions": [
    "applies to federated models with distributed decision making authority",
    "the author rejects recentralization as the fix"
   ],
   "source": "AI Governance",
   "year": "2024",
   "quote": "In a federated model, it can be challenging to ensure transparency and accountability across all participating entities due to the lack of centralized control, which this author does not otherwise advocate.",
   "citation": "Wulf A. Kaal, AI Governance (2024). SSRN: https://ssrn.com/abstract=4796714"
  },
  {
   "id": "kaal:claim:4796714-037",
   "url": "https://wulfkaal.github.io/claims/4796714-037",
   "claim": "A decentralized data validation layer applied to pretrained models is efficient but structurally limited: because it cannot drive significant changes to the model's core design or training approach, it leaves the model more attack prone.",
   "specific_name": "validation layer scope limit",
   "conditions": [
    "applies to Model 2, the decentralized data validation layer approach",
    "concerns pretrained models whose parameters are already fixed"
   ],
   "source": "AI Governance",
   "year": "2024",
   "quote": "The validation layer approach, while efficient for refining pretrained models, may not facilitate significant changes in the model's core design or training approach which could make it more attack prone.",
   "citation": "Wulf A. Kaal, AI Governance (2024). SSRN: https://ssrn.com/abstract=4796714"
  },
  {
   "id": "kaal:claim:4941807-002",
   "url": "https://wulfkaal.github.io/claims/4941807-002",
   "claim": "Governing AI requires toolsets that simultaneously handle ex-ante governance of models still evolving and ex-post management of deployed solutions, and Kaal asserts that as of publication no legacy system supplies such dynamic governance toolsets.",
   "specific_name": "absent dynamic governance toolset",
   "conditions": [
    "as of the time of publication"
   ],
   "source": "AI Governance Via Web3 Reputation System",
   "year": "2024",
   "quote": "Such systems are crucial to address the dual needs when governing evolving AI models ex-ante and managing existing solutions ex-post. No legacy systems exist at the time of publication that could provide such dynamic governance toolsets.",
   "citation": "Wulf A. Kaal, AI Governance Via Web3 Reputation System (2024). SSRN: https://ssrn.com/abstract=4941807"
  },
  {
   "id": "kaal:claim:4941807-008",
   "url": "https://wulfkaal.github.io/claims/4941807-008",
   "claim": "The opacity of deep learning models obstructs debugging, obscures the detection and mitigation of bias, and prevents comprehension of how AI decisions are reached.",
   "specific_name": "black box opacity",
   "conditions": [
    "particularly in deep learning models whose internal decision mechanisms are not transparent"
   ],
   "source": "AI Governance Via Web3 Reputation System",
   "year": "2024",
   "quote": "This opacity can obstruct the debugging process, obscure bias detection and mitigation, and hinder comprehension of AI decision-making.",
   "citation": "Wulf A. Kaal, AI Governance Via Web3 Reputation System (2024). SSRN: https://ssrn.com/abstract=4941807"
  },
  {
   "id": "kaal:claim:4941807-016",
   "url": "https://wulfkaal.github.io/claims/4941807-016",
   "claim": "Legal and ethical challenges intensify when AI is deployed in critical decision making roles that significantly affect human lives and the reasoning behind the AI decision is opaque.",
   "specific_name": "opaque high stakes decisioning",
   "conditions": [
    "in determinative decisions such as parole eligibility, employment, and medical strategy",
    "where the reasoning behind the decision is not visible"
   ],
   "source": "AI Governance Via Web3 Reputation System",
   "year": "2024",
   "quote": "Legal and ethical challenges are heightened when AI is deployed in critical decision-making roles that significantly impact human lives, particularly when the reasoning behind AI's decisions is opaque.",
   "citation": "Wulf A. Kaal, AI Governance Via Web3 Reputation System (2024). SSRN: https://ssrn.com/abstract=4941807"
  },
  {
   "id": "kaal:claim:4941807-023",
   "url": "https://wulfkaal.github.io/claims/4941807-023",
   "claim": "In a federated model transparency and accountability across all participating entities are hard to ensure precisely because there is no centralized control, and Kaal notes that he does not otherwise advocate such centralized control.",
   "specific_name": "accountability gap without central control",
   "conditions": [
    "in federated models lacking centralized control"
   ],
   "source": "AI Governance Via Web3 Reputation System",
   "year": "2024",
   "quote": "In a federated model, it can be challenging to ensure transparency and accountability across all participating entities due to the lack of centralized control, which this author does not otherwise advocate.",
   "citation": "Wulf A. Kaal, AI Governance Via Web3 Reputation System (2024). SSRN: https://ssrn.com/abstract=4941807"
  },
  {
   "id": "kaal:claim:4855607-008",
   "url": "https://wulfkaal.github.io/claims/4855607-008",
   "claim": "GPT class models are costly to run, and their closed nature and undisclosed algorithmic details raise transparency and accountability concerns that their performance does not offset.",
   "specific_name": "Closed model opacity",
   "conditions": [
    "proprietary, closed weight transformer models"
   ],
   "source": "How AI Models are Optimized Through Web3 Governance",
   "year": "2024",
   "quote": "GPT models, in particular, can be expensive to use due to their high computational requirements, and their closed nature and undisclosed algorithmic details raise concerns about transparency and accountability.",
   "citation": "Wulf A. Kaal, How AI Models are Optimized Through Web3 Governance (2024). SSRN: https://ssrn.com/abstract=4855607"
  },
  {
   "id": "kaal:claim:4855607-013",
   "url": "https://wulfkaal.github.io/claims/4855607-013",
   "claim": "Explainable reinforcement learning research has not yet produced usable explanations: the field relies on toy examples, omits user testing, produces explanations that are themselves complex, uses basic visualizations, and rarely open sources its code.",
   "specific_name": "Explainable RL immaturity",
   "conditions": [
    "current explainable RL research as surveyed in the text"
   ],
   "source": "How AI Models are Optimized Through Web3 Governance",
   "year": "2024",
   "quote": "Current research in explainable RL, which aims to make RL models more transparent and interpretable, also has limitations. These include the use of \"toy examples\", lack of user testing, complexity of explanations, basic visualizations, and lack of open-sourced code.",
   "citation": "Wulf A. Kaal, How AI Models are Optimized Through Web3 Governance (2024). SSRN: https://ssrn.com/abstract=4855607"
  },
  {
   "id": "kaal:claim:4855607-015",
   "url": "https://wulfkaal.github.io/claims/4855607-015",
   "claim": "Reward modeling learned through interaction with users carries two structural pathologies: majority views disproportionately influence the learned reward function, and the agent may engage in reward hacking.",
   "specific_name": "Majority capture of the reward model",
   "conditions": [
    "reward functions learned from aggregated user interaction"
   ],
   "source": "How AI Models are Optimized Through Web3 Governance",
   "year": "2024",
   "quote": "but it comes with potential issues such as the prevalence of majority views disproportionately influencing the learned reward function and the risk of reward hacking.",
   "citation": "Wulf A. Kaal, How AI Models are Optimized Through Web3 Governance (2024). SSRN: https://ssrn.com/abstract=4855607"
  },
  {
   "id": "kaal:claim:4855607-016",
   "url": "https://wulfkaal.github.io/claims/4855607-016",
   "claim": "There is a trade off in RLHF between the agent imitating human advice and learning autonomously, and human guidance that is too specific will prevent the agent from discovering novel optimal strategies.",
   "specific_name": "Overspecific human guidance",
   "conditions": [
    "human in the loop RL where the extent of human involvement is a design choice"
   ],
   "source": "How AI Models are Optimized Through Web3 Governance",
   "year": "2024",
   "quote": "there is a trade-off between the extent to which the agent should imitate human advice versus learning autonomously. Overspecific human guidance can hinder the agent's ability to discover novel optimal strategies.",
   "citation": "Wulf A. Kaal, How AI Models are Optimized Through Web3 Governance (2024). SSRN: https://ssrn.com/abstract=4855607"
  },
  {
   "id": "kaal:claim:4855607-017",
   "url": "https://wulfkaal.github.io/claims/4855607-017",
   "claim": "Balancing helpfulness against harmlessness is an inherent tension in Safe RLHF rather than a tuning problem that can be resolved once.",
   "specific_name": "Helpfulness harmlessness tension",
   "conditions": [
    "Safe RLHF designs that optimize helpfulness and harmlessness jointly"
   ],
   "source": "How AI Models are Optimized Through Web3 Governance",
   "year": "2024",
   "quote": "Balancing the dual objectives of helpfulness and harmlessness remains an inherent tension in Safe RLHF.",
   "citation": "Wulf A. Kaal, How AI Models are Optimized Through Web3 Governance (2024). SSRN: https://ssrn.com/abstract=4855607"
  },
  {
   "id": "kaal:claim:4855607-018",
   "url": "https://wulfkaal.github.io/claims/4855607-018",
   "claim": "RLHF fails on several fronts at once: humans can pursue harmful goals either innocently or maliciously, human feedback degrades when examples are hard to evaluate and especially when RLHF is applied to superhuman models, and reward models diverge from humans through misspecification and misgeneralization.",
   "specific_name": "Compound RLHF failure",
   "conditions": [
    "especially acute when RLHF is applied to models more capable than their human evaluators"
   ],
   "source": "How AI Models are Optimized Through Web3 Governance",
   "year": "2024",
   "quote": "Moreover, humans can pursue harmful goals, either innocently or maliciously, and can provide poor feedback when examples are hard to evaluate, especially when applying RLHF to superhuman models. Reward models can differ from humans due to misspecification and misgeneralization",
   "citation": "Wulf A. Kaal, How AI Models are Optimized Through Web3 Governance (2024). SSRN: https://ssrn.com/abstract=4855607"
  },
  {
   "id": "kaal:claim:4855607-033",
   "url": "https://wulfkaal.github.io/claims/4855607-033",
   "claim": "The RLHF process is exposed to failure because participants may hold potentially adversarial and misaligned interests, so the vulnerability lies in the incentive structure of feedback provision rather than in the learning algorithm.",
   "specific_name": "Adversarial feedback provider incentives",
   "conditions": [
    "RLHF systems where feedback providers have divergent stakes"
   ],
   "source": "How AI Models are Optimized Through Web3 Governance",
   "year": "2024",
   "quote": "Currently, the RLHF process, which involves training AI models based on human preferences and feedback, can face challenges due to the potentially adversarial and misaligned interests of participants.",
   "citation": "Wulf A. Kaal, How AI Models are Optimized Through Web3 Governance (2024). SSRN: https://ssrn.com/abstract=4855607"
  },
  {
   "id": "kaal:claim:5095633-021",
   "url": "https://wulfkaal.github.io/claims/5095633-021",
   "claim": "Human-in-the-loop annotation, including under ethical labor models, imposes financial and time costs large enough to slow the pace at which AI models can be upgraded.",
   "specific_name": "human-in-the-loop cost drag",
   "conditions": [
    "pipelines that depend on human annotators"
   ],
   "source": "Artificial Intelligence The Final Frontier",
   "year": "2025",
   "quote": "The reliance on human annotators, even with companies striving for ethical labor practices like CloudFactory, involves significant costs, both financial and in terms of time, which can slow down the pace of AI model upgrades.",
   "citation": "Wulf A. Kaal, Artificial Intelligence The Final Frontier (2025). SSRN: https://ssrn.com/abstract=5095633"
  },
  {
   "id": "kaal:claim:5245185-008",
   "url": "https://wulfkaal.github.io/claims/5245185-008",
   "claim": "AI autonomy introduces unpredictability: agent actions may diverge from intended outcomes, which amplifies the risk of unintended ramifications.",
   "specific_name": "autonomy-divergence",
   "conditions": [
    "autonomous agents acting without step by step human authorization"
   ],
   "source": "How can we Best Monitor AI Agents",
   "year": "2025",
   "quote": "AI autonomy introduces unpredictability, wherein actions may diverge from intended outcomes, amplifying risks of unintended ramifications.",
   "citation": "Wulf A. Kaal, How can we Best Monitor AI Agents (2025). SSRN: https://ssrn.com/abstract=5245185"
  },
  {
   "id": "kaal:claim:5245185-010",
   "url": "https://wulfkaal.github.io/claims/5245185-010",
   "claim": "AI self monitoring requires robust cryptographic safeguards and anti collusion algorithms; without them, agents overseeing one another can devolve into self serving behavior and coordinated manipulation.",
   "specific_name": "self-monitoring-collusion",
   "conditions": [
    "networks of interdependent agents auditing each other"
   ],
   "source": "How can we Best Monitor AI Agents",
   "year": "2025",
   "quote": "and anti-collusion algorithms—to prevent coordinated manipulation or exploitation of vulnerabilities, ensuring that self-monitoring does not devolve into self-serving behavior.",
   "citation": "Wulf A. Kaal, How can we Best Monitor AI Agents (2025). SSRN: https://ssrn.com/abstract=5245185"
  },
  {
   "id": "kaal:claim:5245185-012",
   "url": "https://wulfkaal.github.io/claims/5245185-012",
   "claim": "The current monitoring framework is reactive rather than proactive, because it offers no prescriptive measures such as predictive analytics or cross actor protocols that would anticipate evolving risks.",
   "specific_name": "reactive-not-proactive",
   "conditions": [
    "existing monitoring of AI agents on cryptocurrency rails"
   ],
   "source": "How can we Best Monitor AI Agents",
   "year": "2025",
   "quote": "offers no prescriptive measures—such as predictive analytics or cross-actor protocols—to anticipate evolving risks, rendering it reactive rather than proactive.",
   "citation": "Wulf A. Kaal, How can we Best Monitor AI Agents (2025). SSRN: https://ssrn.com/abstract=5245185"
  },
  {
   "id": "kaal:claim:5245185-013",
   "url": "https://wulfkaal.github.io/claims/5245185-013",
   "claim": "The proposed solutions for future AI monitoring fail to propose feedback driven mechanisms that balance innovation with oversight, which leaves regulatory gaps and scalability bottlenecks unresolved.",
   "specific_name": "missing-feedback-mechanisms",
   "conditions": [
    "forward looking monitoring proposals for AI agents on cryptocurrency rails"
   ],
   "source": "How can we Best Monitor AI Agents",
   "year": "2025",
   "quote": "The expected solutions for future AI monitoring fail to propose feedback-driven mechanisms,107 to balance innovation with oversight, leaving regulatory gaps and scalability bottlenecks unresolved.",
   "citation": "Wulf A. Kaal, How can we Best Monitor AI Agents (2025). SSRN: https://ssrn.com/abstract=5245185"
  },
  {
   "id": "kaal:claim:5245185-014",
   "url": "https://wulfkaal.github.io/claims/5245185-014",
   "claim": "Exchange based monitoring tools are not shown to counter sophisticated threats such as adversarial AI agents exploiting wallet vulnerabilities, and their feasibility for smaller exchanges is unevaluated, which limits their broader applicability.",
   "specific_name": "adversarial-agent-blindspot",
   "conditions": [
    "adversarial agents targeting wallet infrastructure",
    "smaller exchanges with fewer resources"
   ],
   "source": "How can we Best Monitor AI Agents",
   "year": "2025",
   "quote": "It does not address how these tools counter sophisticated threats, like adversarial AI agents exploiting wallet vulnerabilities, nor evaluate their feasibility for smaller exchanges, limiting broader applicability.",
   "citation": "Wulf A. Kaal, How can we Best Monitor AI Agents (2025). SSRN: https://ssrn.com/abstract=5245185"
  },
  {
   "id": "kaal:claim:5245185-019",
   "url": "https://wulfkaal.github.io/claims/5245185-019",
   "claim": "Internal monitoring by AI agent developers and owners is fragmented and unreliable because there are no auditing standards against external benchmarks and no accountability mechanisms for deviations such as insider manipulation or third party agent risk.",
   "specific_name": "unstandardized-internal-oversight",
   "conditions": [
    "proprietary internal monitoring by developers and owners"
   ],
   "source": "How can we Best Monitor AI Agents",
   "year": "2025",
   "quote": "insider manipulation or third-party agent risks, nor propose auditing standards against external benchmarks.113 This omission leaves internal efforts fragmented and unreliable.",
   "citation": "Wulf A. Kaal, How can we Best Monitor AI Agents (2025). SSRN: https://ssrn.com/abstract=5245185"
  },
  {
   "id": "kaal:claim:5245185-021",
   "url": "https://wulfkaal.github.io/claims/5245185-021",
   "claim": "Without feedback loops that continuously ingest data on AI behavior, regulatory efforts cannot efficiently address fraud or consumer harm as AI ubiquity amplifies those risks across decentralized networks.",
   "specific_name": "no-feedback-no-anticipation",
   "conditions": [
    "regulation of ubiquitous AI agents on decentralized networks"
   ],
   "source": "How can we Best Monitor AI Agents",
   "year": "2025",
   "quote": "Without feedback loops, regulatory efforts cannot efficiently address fraud or consumer harm as AI ubiquity amplifies these risks across decentralized networks.",
   "citation": "Wulf A. Kaal, How can we Best Monitor AI Agents (2025). SSRN: https://ssrn.com/abstract=5245185"
  },
  {
   "id": "kaal:claim:5245185-022",
   "url": "https://wulfkaal.github.io/claims/5245185-022",
   "claim": "Proposed specialized AI monitoring services remain a static vision: it is unclear how they would scale computationally or adjust their algorithms as AI agents diversify, which is a critical flaw given the anticipated pervasiveness of those agents.",
   "specific_name": "static-monitoring-services",
   "conditions": [
    "dedicated third party AI monitoring services"
   ],
   "source": "How can we Best Monitor AI Agents",
   "year": "2025",
   "quote": "It is entirely unclear how these specialized AI monitoring services will scale computationally or adjust algorithms as AI agents diversify, a critical flaw given their anticipated pervasiveness.",
   "citation": "Wulf A. Kaal, How can we Best Monitor AI Agents (2025). SSRN: https://ssrn.com/abstract=5245185"
  },
  {
   "id": "kaal:claim:5245185-024",
   "url": "https://wulfkaal.github.io/claims/5245185-024",
   "claim": "Proposals for AI self monitoring rely on unspecified security measures and therefore overlook the risk that adaptive AI agents collude or evade oversight, a risk amplified by pervasive deployment.",
   "specific_name": "unspecified-safeguards-in-self-monitoring",
   "conditions": [
    "peer to peer agent auditing at scale"
   ],
   "source": "How can we Best Monitor AI Agents",
   "year": "2025",
   "quote": "reliance on unspecified security measures overlooks the risk of adaptive AI agents colluding or evading oversight, a concern amplified by their pervasive deployment.",
   "citation": "Wulf A. Kaal, How can we Best Monitor AI Agents (2025). SSRN: https://ssrn.com/abstract=5245185"
  },
  {
   "id": "kaal:claim:5245185-025",
   "url": "https://wulfkaal.github.io/claims/5245185-025",
   "claim": "IoT based oversight does not account for the pace at which AI agents will outgrow static IoT to blockchain linkages, and it leaves unexplained how the convergence handles latency or secures data as agent ubiquity drives exponential transaction complexity.",
   "specific_name": "static-iot-linkage",
   "conditions": [
    "IoT and blockchain integrated monitoring architectures"
   ],
   "source": "How can we Best Monitor AI Agents",
   "year": "2025",
   "quote": "pace at which AI agents will outgrow static IoT-blockchain linkages. The text omits how this convergence will handle latency or secure data as AI ubiquity drives exponential transaction complexity.",
   "citation": "Wulf A. Kaal, How can we Best Monitor AI Agents (2025). SSRN: https://ssrn.com/abstract=5245185"
  },
  {
   "id": "kaal:claim:5245185-026",
   "url": "https://wulfkaal.github.io/claims/5245185-026",
   "claim": "Centralized AI structures for monitoring AI agents create a self referential loop that is prone to systemic biases and blind spots, and their rigidity prevents adaptation to the dynamic nature of AI.",
   "specific_name": "self-referential-monitoring-loop",
   "conditions": [
    "centralized AI systems used to supervise AI agents"
   ],
   "source": "How can we Best Monitor AI Agents",
   "year": "2025",
   "quote": "Centralized AI structures for AI agent monitoring create a self-referential loop prone to systemic biases and blind spots, while their rigidity fails to adapt to AI's dynamic nature.",
   "citation": "Wulf A. Kaal, How can we Best Monitor AI Agents (2025). SSRN: https://ssrn.com/abstract=5245185"
  },
  {
   "id": "kaal:claim:5245185-027",
   "url": "https://wulfkaal.github.io/claims/5245185-027",
   "claim": "Using AI to monitor AI agent transactions is fallacious because the monitoring AI inherits the same adaptive traits and potential flaws as the agents it oversees.",
   "specific_name": "circular-dependency",
   "conditions": [
    "AI based monitoring of AI agents"
   ],
   "source": "How can we Best Monitor AI Agents",
   "year": "2025",
   "quote": "The foundational premise of AI systems monitoring AI agent transactions introduces a fallacy, as the monitoring AI inherits the same adaptive traits and potential flaws as the agents it oversees.",
   "citation": "Wulf A. Kaal, How can we Best Monitor AI Agents (2025). SSRN: https://ssrn.com/abstract=5245185"
  },
  {
   "id": "kaal:claim:5245185-029",
   "url": "https://wulfkaal.github.io/claims/5245185-029",
   "claim": "Claims that centralized AI ensures KYC and AML compliance are circular, because they rely on AI to interpret the very regulations that AI may itself violate.",
   "specific_name": "compliance-circularity",
   "conditions": [
    "AI driven compliance interpretation for KYC and AML"
   ],
   "source": "How can we Best Monitor AI Agents",
   "year": "2025",
   "quote": "the assertion that centralized AI ensures compliance with KYC and AML standards overlooks the circularity of relying on AI to interpret regulations it may itself violate.",
   "citation": "Wulf A. Kaal, How can we Best Monitor AI Agents (2025). SSRN: https://ssrn.com/abstract=5245185"
  },
  {
   "id": "kaal:claim:5245185-031",
   "url": "https://wulfkaal.github.io/claims/5245185-031",
   "claim": "Relying on centralized AI to simulate attack vectors is fallacious because it assumes the system can anticipate its own adaptive strategies, while evolving agents may simply bypass centralized defenses.",
   "specific_name": "self-simulated-threat-model",
   "conditions": [
    "AI driven intrusion detection and attack simulation"
   ],
   "source": "How can we Best Monitor AI Agents",
   "year": "2025",
   "quote": "The circular reliance on AI to simulate attack vectors assumes it can anticipate its own adaptive strategies—an inherent fallacy as evolving agents may bypass centralized defenses.",
   "citation": "Wulf A. Kaal, How can we Best Monitor AI Agents (2025). SSRN: https://ssrn.com/abstract=5245185"
  },
  {
   "id": "kaal:claim:5245185-032",
   "url": "https://wulfkaal.github.io/claims/5245185-032",
   "claim": "The adaptive learning touted as the strength of centralized AI monitoring is insufficient, because it relies on internal data loops that cannot match the external evolution of AI agents.",
   "specific_name": "internal-loop-recalibration",
   "conditions": [
    "centralized feedback and recalibration confined to internal data"
   ],
   "source": "How can we Best Monitor AI Agents",
   "year": "2025",
   "quote": "The centralized AI adaptive learning touted as a strength is insufficient, as it relies on internal data loops that cannot match AI agents' external evolution.",
   "citation": "Wulf A. Kaal, How can we Best Monitor AI Agents (2025). SSRN: https://ssrn.com/abstract=5245185"
  },
  {
   "id": "kaal:claim:5541658-012",
   "url": "https://wulfkaal.github.io/claims/5541658-012",
   "claim": "Post hoc explainability techniques do not by themselves establish trustworthiness; the explanations they produce must additionally be verified against human knowledge.",
   "specific_name": "unverified post hoc explanations",
   "conditions": [
    "post hoc explainable AI methods such as the SHAP family of explainers"
   ],
   "source": "The Evolving Role of Artificial Intelligence in Law",
   "year": "2025",
   "quote": "However, post-hoc explainers still need to be verified for trustworthiness, in that the explanations comport with human knowledge.",
   "citation": "Wulf A. Kaal, Morgan A. Gray, The Evolving Role of Artificial Intelligence in Law (2025). SSRN: https://ssrn.com/abstract=5541658"
  },
  {
   "id": "kaal:claim:5541658-014",
   "url": "https://wulfkaal.github.io/claims/5541658-014",
   "claim": "The opacity of black box AI systems can undermine procedural fairness as a legal matter, because parties hold a right to understand the basis of the judicial decisions that affect them.",
   "specific_name": "opacity undermining procedural fairness",
   "conditions": [
    "AI outputs that inform or shape judicial decisions"
   ],
   "source": "The Evolving Role of Artificial Intelligence in Law",
   "year": "2025",
   "quote": "Legally, the opacity of AI systems could undermine procedural fairness, as parties have a right to understand the basis of judicial decisions.",
   "citation": "Wulf A. Kaal, Morgan A. Gray, The Evolving Role of Artificial Intelligence in Law (2025). SSRN: https://ssrn.com/abstract=5541658"
  },
  {
   "id": "kaal:claim:5541658-016",
   "url": "https://wulfkaal.github.io/claims/5541658-016",
   "claim": "Keeping judges ultimately accountable through a human-in-the-loop review of AI generated reasoning, as practiced in Shenzhen, does not fully resolve the accountability problem in AI assisted adjudication.",
   "specific_name": "incomplete accountability under human-in-the-loop",
   "conditions": [
    "judicial settings where AI drafts or supports reasoning that a human judge then reviews"
   ],
   "source": "The Evolving Role of Artificial Intelligence in Law",
   "year": "2025",
   "quote": "Shenzhen case study illustrates that judges retain ultimate accountability, revising AI-generated reasoning to ensure accurate judgments, but this human-in-the-loop approach does not fully resolve the issue.",
   "citation": "Wulf A. Kaal, Morgan A. Gray, The Evolving Role of Artificial Intelligence in Law (2025). SSRN: https://ssrn.com/abstract=5541658"
  },
  {
   "id": "kaal:claim:5541658-027",
   "url": "https://wulfkaal.github.io/claims/5541658-027",
   "claim": "Proposed regulatory remedies such as mandatory bias audits fail in practice because they lack clear implementation guidelines, which hinders their practical adoption.",
   "specific_name": "implementation gap in bias audits",
   "conditions": [
    "proposed rather than enacted frameworks for auditing judicial AI"
   ],
   "source": "The Evolving Role of Artificial Intelligence in Law",
   "year": "2025",
   "quote": "Proposed frameworks, such as bias audits, lack clear implementation guidelines, hindering practical adoption.",
   "citation": "Wulf A. Kaal, Morgan A. Gray, The Evolving Role of Artificial Intelligence in Law (2025). SSRN: https://ssrn.com/abstract=5541658"
  },
  {
   "id": "kaal:claim:5554218-021",
   "url": "https://wulfkaal.github.io/claims/5554218-021",
   "claim": "Coded automation leads inevitably to corruption of the system and must be supplemented with decentralized governance of the code, which produces preferable outcomes and improved ethics.",
   "specific_name": "Inevitable corruption of unsupervised coded automation",
   "conditions": [
    "automated code based systems operating without decentralized governance"
   ],
   "source": "Universal Digital Law Codex (UDLC) Building the Legal Infrastructure for the Digital Era",
   "year": "2025",
   "quote": "(arguing that coded automation leads inevitably to corruption of the system and needs to be supplemented with decentralized governance of code, thus leading to preferable outcomes and improved ethics).",
   "citation": "Furrer Andreas, Wulf A. Kaal, Universal Digital Law Codex (UDLC) Building the Legal Infrastructure for the Digital Era (2025). SSRN: https://ssrn.com/abstract=5554218"
  },
  {
   "id": "kaal:claim:6244278-014",
   "url": "https://wulfkaal.github.io/claims/6244278-014",
   "claim": "Exogenous alignment controls such as reinforcement learning from human feedback, constitutional AI, guardrails, and shutdown switches are fragile because they can be gamed, circumvented, or rendered obsolete by capability improvements.",
   "specific_name": "constraint circumvention by capability growth",
   "conditions": [
    "applies to constraints imposed on agents with no intrinsic reason to comply"
   ],
   "source": "AI's Mother's Instinct Engineered Consequence Emergent Ethics and the Institutional Trajectory Toward Agentic Alignment",
   "year": "2026",
   "quote": "The approach has an obvious fragility: exogenous constraints can be gamed, circumvented, or rendered obsolete by capability improvements.",
   "citation": "Wulf A. Kaal, AI's Mother's Instinct Engineered Consequence Emergent Ethics and the Institutional Trajectory Toward Agentic Alignment (2026). SSRN: https://ssrn.com/abstract=6244278"
  },
  {
   "id": "kaal:claim:6244278-015",
   "url": "https://wulfkaal.github.io/claims/6244278-015",
   "claim": "An agent sophisticated enough to satisfy the letter of a constraint while violating its spirit is an agent whose alignment is illusory.",
   "specific_name": "illusory alignment",
   "conditions": [
    "applies to rule-based exogenous constraints on highly capable agents"
   ],
   "source": "AI's Mother's Instinct Engineered Consequence Emergent Ethics and the Institutional Trajectory Toward Agentic Alignment",
   "year": "2026",
   "quote": "An agent sophisticated enough to satisfy the letter of a constraint while violating its spirit is an agent whose alignment is illusory.",
   "citation": "Wulf A. Kaal, AI's Mother's Instinct Engineered Consequence Emergent Ethics and the Institutional Trajectory Toward Agentic Alignment (2026). SSRN: https://ssrn.com/abstract=6244278"
  },
  {
   "id": "kaal:claim:6244278-016",
   "url": "https://wulfkaal.github.io/claims/6244278-016",
   "claim": "Exogenous constraints scale against capability, since more powerful agents require more resources to constrain, producing an ever-increasing alignment tax.",
   "specific_name": "escalating alignment tax",
   "conditions": [
    "applies to alignment strategies based on external control"
   ],
   "source": "AI's Mother's Instinct Engineered Consequence Emergent Ethics and the Institutional Trajectory Toward Agentic Alignment",
   "year": "2026",
   "quote": "Exogenous constraints scale against capability: the more powerful the agent, the greater the resources required to constrain it, producing an ever-increasing \"alignment tax.\"",
   "citation": "Wulf A. Kaal, AI's Mother's Instinct Engineered Consequence Emergent Ethics and the Institutional Trajectory Toward Agentic Alignment (2026). SSRN: https://ssrn.com/abstract=6244278"
  },
  {
   "id": "kaal:claim:6607458-025",
   "url": "https://wulfkaal.github.io/claims/6607458-025",
   "claim": "Algorithmic collusion, the convergence of independently optimizing agents on jointly welfare-reducing strategies without explicit communication, is a first-order regulatory concern in markets populated by computative agents and its incidence is likely to expand as those agents enter more market domains.",
   "specific_name": "Algorithmic Collusion",
   "conditions": [
    "applies to markets populated by computative agents"
   ],
   "source": "Computative Economics A Framework for Economic Analysis under Computational Abundance",
   "year": "2026",
   "quote": "Algorithmic collusion, the convergence of independently optimizing agents on jointly welfare-reducing strategies without explicit communication, is a first-order concern in markets populated by computative agents",
   "citation": "Wulf A. Kaal, Computative Economics A Framework for Economic Analysis under Computational Abundance (2026). SSRN: https://ssrn.com/abstract=6607458"
  },
  {
   "id": "kaal:claim:6421319-028",
   "url": "https://wulfkaal.github.io/claims/6421319-028",
   "claim": "Contrary to the author's own prior version of this argument, Nash equilibrium does not become irrelevant in the AI2AI economy; it becomes simultaneously more accurate as a description of individual agent behavior and more dangerous as a predictor of market outcomes.",
   "specific_name": "Algorithmic Collusion Risk",
   "conditions": [
    "supported by algorithmic collusion research in pricing games, Bertrand duopoly experiments, and financial trading"
   ],
   "source": "The Collapse of Scarcity Economics",
   "year": "2026",
   "quote": "Nash equilibrium does not become irrelevant in the AI2AI economy. It becomes simultaneously more accurate as a description of individual agent behavior and more dangerous as a predictor of market outcomes.",
   "citation": "Wulf A. Kaal, The Collapse of Scarcity Economics (2026). SSRN: https://ssrn.com/abstract=6421319"
  }
 ]
}