# Propositions from Catalytic Regulation: Incentivizing Safety During a Regulatory Drought

**Citation:** Yonathan A. Arbel, Catalytic Regulation: Incentivizing Safety During a Regulatory Drought, AI Frontiers (2026).

**Source:** [unpaginated AI Frontiers online essay](https://ai-frontiers.org/articles/catalytic-regulation-incentivizing-safety-during-a-regulatory-drought)

**Review status:** 22 model-drafted, source-checked; 0 human-reviewed. This online-only work uses section-level unpaginated anchors.

## 1. AI safety needs a Volvo moment in which a safety innovation becomes a source of market leadership rather than a competitive handicap

**Location:** Opening, The Volvo Analogy, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, in the opening, that Volvo’s decision to make the three-point seat belt broadly available showed that safety could define excellence even when competitors believed “safety doesn’t sell.” American AI needs an analogous moment in which reliable safety becomes a recognized basis of competition. This is significant because the proposal begins from market identity and consumer value, not only from fear of legal punishment. It connects to Volvo, the three-point seat belt, safety innovation, brand reputation, market leadership, and AI safety culture.

**Evidence anchor:** The essay contrasts Ford’s unsuccessful surcharge-based safety campaign with Volvo’s safety-led identity and calls for an American AI “Volvo moment.”

**Boundary:** The automobile analogy is illustrative; AI safety lacks a single, proven intervention as clear and measurable as the three-point seat belt.

**Connections:** Volvo; three-point seat belt; safety innovation; brand reputation; market leadership; AI safety culture

**Record:** `catalytic-regulation-p01` · `machine-drafted-source-checked`

## 2. Catalytic regulation uses public rewards to redirect market competition toward safety rather than relying primarily on mandates and penalties

**Location:** Opening, Definition of Catalytic Regulation, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, in the opening, that catalytic regulation is a family of positive incentives. Where command-and-control regulation says “do this or else,” catalytic tools offer tax credits, procurement advantages, market guarantees, or prestige for verified safety work. This is significant because government can shape the direction of innovation without making compliance the sole reason firms invest in safety. It connects to incentive regulation, tax expenditures, procurement, market design, prestige, and safety-by-design.

**Evidence anchor:** The opening defines catalytic regulation by contrasting rewards with mandates and names tax, procurement, and prestige mechanisms.

**Boundary:** Positive incentives do not guarantee universal adoption and are presented as complements to, not replacements for, every mandate or liability rule.

**Connections:** incentive regulation; tax expenditures; procurement; market design; prestige; safety-by-design

**Record:** `catalytic-regulation-p02` · `machine-drafted-source-checked`

## 3. The deeper objective is to make safe and powerful AI a competitive Schelling point for American laboratories

**Location:** Opening, Cultural Objective, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, in the opening, that marginal subsidies are worthwhile but insufficient as the ultimate ambition. Catalytic regulation should coordinate labs, buyers, workers, and policymakers around the idea that American leadership means producing systems that are powerful and demonstrably safe. This is significant because a visible focal point can change what organizations compete to optimize even after a particular payment ends. It connects to Schelling points, norm formation, competitive equilibrium, organizational identity, American AI leadership, and cultural change.

**Evidence anchor:** The essay says the goal is to catalyze a cultural shift that defines safe and powerful AI as the distinctive strength of American labs.

**Boundary:** Whether safety can become a stable focal point depends on credible measurement, visible rewards, and continued market or professional demand.

**Connections:** Schelling points; norm formation; competitive equilibrium; organizational identity; American AI leadership; cultural change

**Record:** `catalytic-regulation-p03` · `machine-drafted-source-checked`

## 4. Positive AI-safety incentives can work with current deregulatory politics when traditional mandates are politically blocked

**Location:** Why Catalytic Regulation?, Political Feasibility, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “Why Catalytic Regulation?,” that U.S.–China race rhetoric and organized industry resistance have created a regulatory drought for mandates, prohibitions, and prescriptive standards. Tax credits and procurement incentives can instead be framed as investments in American leadership and do not require legislators to defend slowing domestic firms. This is significant because feasible second-best tools can build safety capacity during a period when waiting for ideal legislation would produce inaction. It connects to regulatory drought, political feasibility, AI race rhetoric, command-and-control regulation, industrial policy, and second-best governance.

**Evidence anchor:** The essay describes federal deregulatory policy, industry opposition, and the politically attractive framing of positive incentives as investment in safer U.S. leadership.

**Boundary:** Working with deregulatory attitudes may preserve political feasibility but can also limit the scale, conditionality, or enforceability of the intervention.

**Connections:** regulatory drought; political feasibility; AI race rhetoric; command-and-control regulation; industrial policy; second-best governance

**Record:** `catalytic-regulation-p04` · `machine-drafted-source-checked`

## 5. Catalytic tools can generate the knowledge, benchmarks, and institutional capacity needed for stronger future regulation

**Location:** Why Catalytic Regulation?, Complementarity, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “Why Catalytic Regulation?,” that uncertainty about effective AI-safety standards is itself a reason to fund experimentation now. Rewards can advance red teaming, model-on-model audits, safety practices, benchmarking infrastructure, and expert capacity that later legislators or regulators can convert into mandates. This is significant because positive incentives are not merely a consolation prize during political stalemate; they can improve the information environment for future law. It connects to regulatory learning, benchmark development, red teaming, institutional capacity, experimental governance, and dynamic regulation.

**Evidence anchor:** The essay argues that present incentives can develop safety techniques and infrastructure that make later mandates more informed and effective.

**Boundary:** Knowledge produced under subsidy may remain incomplete, proprietary, or strategically reported unless programs include disclosure and verification requirements.

**Connections:** regulatory learning; benchmark development; red teaming; institutional capacity; experimental governance; dynamic regulation

**Record:** `catalytic-regulation-p05` · `machine-drafted-source-checked`

## 6. Capability-correlated safety benchmarks can mislabel general capability gains as safety progress and complicate standard setting

**Location:** Why Catalytic Regulation?, Measurement Limits, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “Why Catalytic Regulation?,” that many purported safety benchmarks rise with model capability. A lab may therefore appear safer because its system is generally stronger, even without a targeted reduction in dangerous behavior or failure modes. This is significant because incentives must distinguish genuine safety investment from capability work or they will subsidize the race they are meant to redirect. It connects to benchmark validity, Goodhart’s law, capability-safety correlation, program integrity, measurement error, and regulatory standards.

**Evidence anchor:** The article notes empirical correlation between capability and safety benchmarks and warns that capability progress can be portrayed as safety improvement.

**Boundary:** The essay identifies the confound and points to attestation and audits in companion work but does not supply a complete classification rule here.

**Connections:** benchmark validity; Goodhart’s law; capability-safety correlation; program integrity; measurement error; regulatory standards

**Record:** `catalytic-regulation-p06` · `machine-drafted-source-checked`

## 7. Rewarded best practices can become evidence of reasonable care and later anchor negligence or regulatory standards

**Location:** Why Catalytic Regulation?, Legal Infrastructure, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “Why Catalytic Regulation?,” that positive incentives may cause leading firms to develop and disclose practices that others recognize as competent safety behavior. Even without universal adoption, those practices can inform the negligence standard of reasonable care and give future regulators concrete reference points. This is significant because a voluntary reward program can acquire indirect legal force through the evolution of professional custom and adjudicative baselines. It connects to reasonable care, negligence, industry custom, soft law, regulatory bootstrapping, and standards development.

**Evidence anchor:** The essay states that catalytic best practices can anchor reasonable care in negligence and guide later regulatory action.

**Boundary:** Courts need not treat subsidized practices as dispositive, and industry custom can be inadequate or strategically shaped.

**Connections:** reasonable care; negligence; industry custom; soft law; regulatory bootstrapping; standards development

**Record:** `catalytic-regulation-p07` · `machine-drafted-source-checked`

## 8. AI safety depends on organizational culture because local knowledge and unobserved decisions exceed regulators’ monitoring capacity

**Location:** Why Catalytic Regulation?, Organizational Culture, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “Why Catalytic Regulation?,” that organizational culture governs what workers do “when no one is looking.” AI laboratories contain countless pockets of technical and operational knowledge that Washington regulators cannot continuously observe, so internal willingness to test, report, and question matters alongside formal compliance. This is significant because rules aimed only at observable outputs miss the distributed decisions through which high-risk systems are built. It connects to organizational culture, local knowledge, monitoring limits, whistleblowing, safety reporting, and management-based regulation.

**Evidence anchor:** The essay uses examples from food service, NASA, and medicine to illustrate hidden choices and applies the lesson to local knowledge inside AI labs.

**Boundary:** Culture is difficult to measure and can be invoked performatively; incentives require observable proxies and independent verification.

**Connections:** organizational culture; local knowledge; monitoring limits; whistleblowing; safety reporting; management-based regulation

**Record:** `catalytic-regulation-p08` · `machine-drafted-source-checked`

## 9. Safety culture can precede mandatory regulation, as high-reliability organizations and historical self-regulation demonstrate

**Location:** Why Catalytic Regulation?, High Reliability Organizations, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “Why Catalytic Regulation?,” that nuclear facilities, air-traffic control, and aircraft carriers manage inherently risky technologies partly through organizational control of hazards and probabilities. The nuclear industry’s post–Three Mile Island self-regulatory body and Volvo’s early seat-belt standard show that safety norms sometimes arise before legal codification. This is significant because the absence of a mandate need not imply waiting passively for disaster or legislative consensus. It connects to high-reliability organizations, nuclear safety, self-regulation, Three Mile Island, professional norms, and anticipatory governance.

**Evidence anchor:** The essay draws lessons from HRO scholarship, the Institute of Nuclear Power Operations, and Volvo’s adoption of the seat belt before federal mandates.

**Boundary:** High-reliability sectors have distinctive missions, oversight, and professional structures that AI labs may not naturally share.

**Connections:** high-reliability organizations; nuclear safety; self-regulation; Three Mile Island; professional norms; anticipatory governance

**Record:** `catalytic-regulation-p09` · `machine-drafted-source-checked`

## 10. Catalytic regulation has positive expected value because low-cost, compatible tools build useful safety scaffolding even when they solve only part of the problem

**Location:** Why Catalytic Regulation?, Marginal Case, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “Why Catalytic Regulation?,” that the relevant marginal question is whether society is better off with feasible positive incentives than without them. These tools are comparatively low cost, politically available, and compatible with other regulatory strategies; they need not solve every safety problem to justify adoption. This is significant because demanding a complete solution can become an excuse for rejecting useful institutional increments. It connects to expected-value reasoning, marginal analysis, policy complementarity, institutional scaffolding, regulatory optionality, and incrementalism.

**Evidence anchor:** The essay argues that catalytic tools foreclose no other strategy and can build scaffolding even when their immediate safety effect is incomplete.

**Boundary:** Low downside is a claim that depends on program design; subsidies can waste funds, entrench incumbents, or reward cosmetic compliance.

**Connections:** expected-value reasoning; marginal analysis; policy complementarity; institutional scaffolding; regulatory optionality; incrementalism

**Record:** `catalytic-regulation-p10` · `machine-drafted-source-checked`

## 11. A practical catalytic-regulation bundle should combine corporate incentives, demand-side incentives, market guarantees, and prestige

**Location:** From Principles to Practice: The Catalytic Regulation Bundle, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “From Principles to Practice: The Catalytic Regulation Bundle,” that no single reward channel is sufficient. The proposed portfolio acts on laboratories’ investment choices, buyers’ demand, the market value of shared innovations, and the status motives of organizations and researchers. This is significant because the four mechanisms address different bottlenecks and can reinforce one another across the safety ecosystem. It connects to policy portfolios, supply-side incentives, demand-side incentives, procurement, intellectual property, and prestige competition.

**Evidence anchor:** The practice section explicitly introduces four mechanisms: corporate incentives, demand-side incentives, market guarantees, and prestige.

**Boundary:** The essay offers illustrative mechanisms rather than a fully costed legislative package or an empirical estimate of interaction effects.

**Connections:** policy portfolios; supply-side incentives; demand-side incentives; procurement; intellectual property; prestige competition

**Record:** `catalytic-regulation-p11` · `machine-drafted-source-checked`

## 12. A safety R&D tax credit would solve an allocation problem inside well-capitalized laboratories rather than a lack of total financing

**Location:** Corporate Incentives, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “Corporate Incentives,” that frontier labs may possess ample capital while allocating too little of the next engineering team, GPU cluster, or research dollar to safety. An earmarked credit changes the relative return to that marginal allocation without attempting to fund the laboratory as a whole. This is significant because observed wealth does not eliminate the case for incentives when private returns steer resources toward capability rather than public safety. It connects to R&D tax credits, marginal allocation, opportunity cost, frontier laboratories, compute budgets, and safety-by-design.

**Evidence anchor:** The corporate-incentive section contrasts total financing with internal allocation and describes how a credit changes the next marginal staffing or compute decision.

**Boundary:** The program must define eligible safety research and avoid subsidizing capability projects that firms would fund anyway.

**Connections:** R&D tax credits; marginal allocation; opportunity cost; frontier laboratories; compute budgets; safety-by-design

**Record:** `catalytic-regulation-p12` · `machine-drafted-source-checked`

## 13. The Orphan Drug Act illustrates how tax incentives can redirect sophisticated firms toward socially valuable research neglected by ordinary market returns

**Location:** Corporate Incentives, Orphan Drug Blueprint, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “Corporate Incentives,” that the Orphan Drug Act helped create a therapeutic category by rewarding research that well-resourced pharmaceutical companies otherwise underprovided. The analogy is not that AI labs lack money but that safety work, like rare-disease research, receives too little of the marginal investment. This is significant because an established innovation-policy instrument offers a blueprint for changing research direction without prescribing technical methods. It connects to the Orphan Drug Act, pharmaceutical innovation, tax policy, underinvestment, positive externalities, and technology-neutral incentives.

**Evidence anchor:** The essay describes the orphan-drug credit’s role in redirecting pharmaceutical research and maps its allocation logic onto AI safety.

**Boundary:** Orphan-drug policy has distinct markets and controversies, so its reported output growth does not prove identical results for AI safety.

**Connections:** Orphan Drug Act; pharmaceutical innovation; tax policy; underinvestment; positive externalities; technology-neutral incentives

**Record:** `catalytic-regulation-p13` · `machine-drafted-source-checked`

## 14. Safety research subsidies can scale with capability progress but require attestation and audit to separate genuine safety work from general development

**Location:** Corporate Incentives, Scalability and Integrity, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “Corporate Incentives,” that policymakers can expand a safety credit when frontier progress accelerates and contract it when development slows. The central integrity problem is classification: general capability work can masquerade as safety work, especially when benchmark performance overlaps. This is significant because an adaptive incentive is only as credible as its eligibility, attestation, and audit architecture. It connects to scalable subsidies, attestation, independent audits, program integrity, capability-safety distinction, and adaptive policy.

**Evidence anchor:** The corporate-incentive section identifies scalability, acknowledges the safety-capability classification problem, and points to attestation and audit mechanisms.

**Boundary:** The essay refers readers to companion work for detailed safeguards and relies partly on reputational blowback to deter elaborate gaming.

**Connections:** scalable subsidies; attestation; independent audits; program integrity; capability-safety distinction; adaptive policy

**Record:** `catalytic-regulation-p14` · `machine-drafted-source-checked`

## 15. Consumer tax credits for certified-safe AI can transmit U.S. standards to foreign developers through market access

**Location:** Demand-Side Incentives, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “Demand-Side Incentives,” that direct jurisdiction over foreign labs is limited but American purchasing power is not. An “AI Safety Usage Credit” could give businesses and individuals a price advantage for buying certified services, inducing foreign providers to qualify for the U.S. market much as energy-efficiency incentives affected global manufacturing. This is significant because domestic demand policy can export a standard without an international treaty or extraterritorial command. It connects to consumption subsidies, market access, extraterritorial influence, energy-efficiency credits, certification, and the Brussels effect.

**Evidence anchor:** The essay uses energy-efficiency incentives and foreign manufacturers’ adaptation to propose a usage credit for certified AI services.

**Boundary:** The analogy depends on the size and attractiveness of the U.S. market and on certification criteria that foreign providers can verify and satisfy.

**Connections:** consumption subsidies; market access; extraterritorial influence; energy-efficiency credits; certification; Brussels effect

**Record:** `catalytic-regulation-p15` · `machine-drafted-source-checked`

## 16. AI safety certification should begin with defensible properties and update as knowledge develops rather than wait for a complete standard

**Location:** Demand-Side Incentives, Certification, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “Demand-Side Incentives,” that certification can initially target properties markets underprice, including adversarial robustness, resistance to deceptive alignment, interpretability of core reasoning, and safeguards against autonomous misuse. Criteria can then be revised semiannually or annually as safety innovation advances. This is significant because uncertainty becomes a reason for iterative standard setting rather than indefinite delay. It connects to adaptive certification, adversarial robustness, deceptive alignment, interpretability, autonomous misuse, and regulatory iteration.

**Evidence anchor:** The demand-side section lists candidate certification properties and proposes periodic revision as safety knowledge matures.

**Boundary:** Early criteria will be incomplete and may invite gaming or false assurance; updating them requires a capable, legitimate certifying institution.

**Connections:** adaptive certification; adversarial robustness; deceptive alignment; interpretability; autonomous misuse; regulatory iteration

**Record:** `catalytic-regulation-p16` · `machine-drafted-source-checked`

## 17. An AI Safety Charter can use federal purchasing and access benefits to guarantee a market reward for demonstrated safety leadership

**Location:** Market Guarantees, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “Market Guarantees,” that tax incentives do not fully solve the problem faced by firms that invest in safety whose benefits spill to competitors. Modeled partly on Operation Warp Speed’s use of purchasing power, an AI Safety Charter could offer procurement priority, faster clearances, streamlined compliance, and research-data access to qualifying firms. This is significant because government can make safety investment privately bankable by attaching a valuable market to verified performance. It connects to federal procurement, Operation Warp Speed, advance market commitments, security clearances, government datasets, and appropriability.

**Evidence anchor:** The market-guarantee section identifies the appropriability problem and proposes a charter bundling procurement and government-access benefits.

**Boundary:** Preferential access can favor incumbents or become politicized unless qualification is transparent, contestable, and independently verified.

**Connections:** federal procurement; Operation Warp Speed; advance market commitments; security clearances; government datasets; appropriability

**Record:** `catalytic-regulation-p17` · `machine-drafted-source-checked`

## 18. FRAND licensing can preserve a temporary private return to safety inventions while preventing patents from bottlenecking diffusion

**Location:** Market Guarantees, Safety Patents, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “Market Guarantees,” that firms may hide breakthrough interpretability or safety methods because publication lets rivals copy them. Patents combined with mandatory fair, reasonable, and nondiscriminatory licensing at modest royalties could reward innovation while spreading the technique broadly. This is significant because the design seeks to reconcile appropriability with the social need for rapid diffusion of protective knowledge. It connects to patents, FRAND licensing, interpretability, knowledge diffusion, innovation incentives, and standard-essential technology.

**Evidence anchor:** The essay proposes modest-royalty FRAND licensing to give safety innovators a return without permitting exclusive bottlenecks.

**Boundary:** Patentability, disclosure quality, royalty setting, and administration could create delay or strategic litigation not addressed in the short essay.

**Connections:** patents; FRAND licensing; interpretability; knowledge diffusion; innovation incentives; standard-essential technology

**Record:** `catalytic-regulation-p18` · `machine-drafted-source-checked`

## 19. Status competition can motivate persistent organizational improvement even without a cash prize

**Location:** The Power of Prestige, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “The Power of Prestige,” that professional standing is a durable motivator distinct from direct payment. The Malcolm Baldrige National Quality Award and OSHA’s Voluntary Protection Programs illustrate how visible recognition can induce process improvement, lower injuries, and change what engineers optimize for. This is significant because AI researchers and labs already compete intensely for citations, rankings, and talent, making prestige a compatible regulatory currency. It connects to status competition, Baldrige Award, OSHA VPP, professional recognition, process improvement, and behavioral incentives.

**Evidence anchor:** The prestige section describes noncash quality and workplace-safety programs and maps their motivational logic onto AI’s ranking culture.

**Boundary:** The reported program outcomes are contextual and may reflect selection effects; prestige works only if the recognizing institution is credible.

**Connections:** status competition; Baldrige Award; OSHA VPP; professional recognition; process improvement; behavioral incentives

**Record:** `catalytic-regulation-p19` · `machine-drafted-source-checked`

## 20. A scarce presidential safety medal and public technical leaderboard could make verified safety achievement visible, prestigious, and talent-attracting

**Location:** The Power of Prestige, Proposed Instruments, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “The Power of Prestige,” that government could recognize a small number of annual safety leaders through White House ceremonies conditioned on publication of practices or tools. An attestation body could separately maintain a public leaderboard for recognized safety challenges, allowing technically credible achievement to influence reputation and recruitment. This is significant because scarcity, publicity, and peer verifiability can make safety status harder to fabricate than a tax claim. It connects to presidential medals, public leaderboards, attestation bodies, talent markets, open safety tools, and reputational incentives.

**Evidence anchor:** The essay proposes a deliberately scarce presidential medal, publication requirements, and an attested leaderboard tied to safety challenges.

**Boundary:** Technical achievements can still be benchmark-gamed, and political recognition risks favoritism unless criteria and adjudication are independent and transparent.

**Connections:** presidential medals; public leaderboards; attestation bodies; talent markets; open safety tools; reputational incentives

**Record:** `catalytic-regulation-p20` · `machine-drafted-source-checked`

## 21. The durable aim is to shift the competitive equilibrium so leading labs define safety as excellence and smaller labs imitate them

**Location:** The Power of Prestige, Equilibrium Shift, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “The Power of Prestige,” that government cannot and need not outspend frontier laboratories. Financial rewards supply activation energy, while visible safety competition changes norms: leaders make safety part of excellence, aspirational firms copy them, and internal safety factions gain influence. This is significant because the same competitive energy that currently accelerates risk-taking can be redirected toward risk reduction. It connects to equilibrium selection, norm cascades, organizational imitation, activation energy, internal coalitions, and competitive safety.

**Evidence anchor:** The essay concludes the prestige discussion by distinguishing modest financial activation from deeper competitive and cultural diffusion across labs.

**Boundary:** The multiplier is a theory of cultural diffusion that requires empirical testing and may fail if capability prestige continues to dominate safety prestige.

**Connections:** equilibrium selection; norm cascades; organizational imitation; activation energy; internal coalitions; competitive safety

**Record:** `catalytic-regulation-p21` · `machine-drafted-source-checked`

## 22. At minimum catalytic regulation seeds future governance; at best it redefines American AI leadership around safety guarantees valued worldwide

**Location:** A Seat Belt for AI, unpaginated online source

Professor Yonathan A. Arbel claims, in “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought,” an unpaginated AI Frontiers essay, under “A Seat Belt for AI,” that positive incentives offer a path during intense skepticism of traditional regulation. Their modest effect is to generate safety advances and institutions for the next regulatory architecture; their ambitious effect is to make safety a nonnegotiable reputational priority that attracts governments, enterprises, and users globally. This is significant because the proposal is robust to uncertain ambition: useful groundwork does not depend on immediately achieving cultural transformation. It connects to robust policy, regulatory sequencing, safety guarantees, global demand, reputational priority, and the Volvo analogy.

**Evidence anchor:** The final section presents catalytic regulation as both groundwork for future architecture and a possible redefinition of the AI race around trusted safety.

**Boundary:** The essay does not establish that policymakers will enact the incentives or that another country could not seize the same safety-leadership strategy first.

**Connections:** robust policy; regulatory sequencing; safety guarantees; global demand; reputational priority; Volvo analogy

**Record:** `catalytic-regulation-p22` · `machine-drafted-source-checked`