Belief: The United States Should Establish a Federal Regulatory Framework for Autonomous Vehicles
Topic: Transportation & Infrastructure > Vehicle Technology > Autonomous Vehicles
Topic IDs: Dewey: 388.3
Belief Positivity Towards Topic: +58%
Claim Magnitude: 68% (Significant structural policy claim affecting transportation safety, industry development, and urban planning. The technology is real and deployable now (Waymo, GM Cruise) but not yet ubiquitous — making this a near-term regulatory design question, not a speculative one. Contested primarily on the prescriptiveness of the framework: most stakeholders agree some federal framework is needed; the dispute is whether it should be permissive or safety-focused.)
Each section builds a complete analysis from multiple angles. View the full technical documentation on GitHub. Created 2026-03-22: Full ISE template population, all 17 sections.
In August 2023, a GM Cruise robotaxi struck a pedestrian, dragged her 20 feet, and then drove forward over her — and the company initially filed inaccurate incident reports with regulators. California's DMV suspended Cruise's operating permit. GM shut down the entire Cruise program. The incident exposed the central problem with the current U.S. approach to autonomous vehicle regulation: there is no federal pre-market safety approval requirement for AV systems. Companies deploy on public roads, people get hurt, and regulators respond after the fact with the equivalent of a recall — except the "product" is a 4,000-lb vehicle moving through public space.
The ISE framing: this is not a debate about whether autonomous vehicles are possible or desirable. Waymo has logged millions of commercial miles in Phoenix and San Francisco with a documented safety record. The question is: who sets the safety bar before these systems are deployed at scale, and how prescriptive should those standards be? That question involves a genuine tension — too permissive and you get Cruise incidents; too prescriptive and you lock in 2024 technology standards that become obstacles to 2030 improvements. The SELF DRIVE Act (2017) and the AV START Act (2017) failed to pass because Congress couldn't resolve that tension. It still hasn't.
📚 Definition of Terms
| Term | Definition as Used in This Belief |
|---|---|
| Autonomous Vehicle (AV) / Self-Driving Vehicle | A vehicle capable of performing the dynamic driving task without ongoing human input. In this belief, "autonomous vehicle" refers primarily to SAE Level 3 (conditional automation — human must resume control when system requests), Level 4 (high automation — no human required within defined operational design domain), and Level 5 (full automation — no human required in any conditions). Level 1 and Level 2 (driver assistance systems like Tesla Autopilot, Cadillac Super Cruise) are partially automated but require continuous driver engagement — they are regulated under existing Federal Motor Vehicle Safety Standards (FMVSS) and are not the primary subject of this belief. The regulatory gap is primarily at Level 3 and above, where the question of when and whether a human must be present changes the applicable safety standard. |
| Operational Design Domain (ODD) | The specific conditions under which an AV system is designed to function: geographic area (geofenced city blocks, highway, etc.), weather conditions, time of day, speed range, and other operating parameters. A robotaxi (Waymo, Cruise) operates in a defined ODD (specific city zones) under Level 4 automation. A delivery vehicle (Nuro, Starship) operates in a defined ODD (low-speed suburban streets). The ODD matters for regulation because it defines the envelope of claims the manufacturer is making about system capabilities — and therefore the safety testing scope that should be required before deployment. Regulatory standards that don't define testing requirements relative to ODD allow manufacturers to self-define their operating envelope after deployment. |
| NHTSA (National Highway Traffic Safety Administration) | The federal agency with authority over motor vehicle safety standards. NHTSA's existing authority is primarily post-market: it can recall vehicles after demonstrating a safety defect, and it can set Federal Motor Vehicle Safety Standards (FMVSS) for vehicle hardware (brakes, lights, crashworthiness). NHTSA lacks pre-market approval authority for AV software — it cannot require a manufacturer to demonstrate safety before deploying an AV system on public roads. This is the primary regulatory gap. The SELF DRIVE Act (2017) and FAV FORWARD Act (2021) would have granted NHTSA limited pre-market authority; both failed. |
| Federal Motor Vehicle Safety Standards (FMVSS) | The set of performance and design standards for motor vehicles, established under the National Traffic and Motor Vehicle Safety Act (1966). FMVSS covers hardware standards — brakes, lighting, crashworthiness, tire pressure monitoring — but was designed for human-driven vehicles. AV systems create problems for FMVSS because: (1) some FMVSS standards assume a human driver (e.g., steering wheel requirements); (2) AV safety is primarily determined by software behavior, which FMVSS does not cover; (3) ODD limitations mean AV safety is conditional on operating context, while FMVSS standards are generally unconditional. NHTSA has granted exemptions from FMVSS requirements (e.g., steering wheel exemptions for autonomous delivery vehicles) on an ad hoc basis, creating inconsistent standards. |
| SAE Level (J3016 Standard) | Society of Automotive Engineers International standard J3016 defines six levels of driving automation: Level 0 (no automation), Level 1 (driver assistance — one function automated, e.g., adaptive cruise control), Level 2 (partial automation — multiple functions automated but driver must remain engaged, e.g., Tesla Autopilot), Level 3 (conditional automation — system handles all driving in specific conditions, but driver must resume when system requests), Level 4 (high automation — system handles all driving in ODD without requiring driver resumption), Level 5 (full automation — no conditions or limitations). The SAE level taxonomy is universally referenced but does not itself specify safety requirements for any level — it is a descriptive classification, not a regulatory standard. |
| Tort Liability / Product Liability | The legal framework for compensating persons harmed by defective products. In AV crashes, existing product liability law (strict liability for manufacturing defects, negligence standard for design defects) faces challenges: (1) Was the crash caused by a software "design defect" or by a scenario outside the ODD (operator error)? (2) Courts are split on whether software is a "product" (subject to strict liability) or a "service" (subject to negligence). (3) When multiple parties are involved (AV manufacturer, sensor manufacturer, mapping data provider, fleet operator, passenger), standard tort apportionment rules produce complex multi-party litigation. The liability uncertainty is both a potential safety incentive (manufacturers bear tort costs of accidents) and a barrier to insurance market development (insurers cannot price risk without statistical history). |
🔍 Argument Trees
Each reason is a belief with its own page. Scoring is recursive based on truth, linkage, and importance.
✅ Top Scoring Reasons to Agree | Argument Score | Linkage Score | Impact |
|---|---|---|---|
| The current regulatory vacuum means AV companies are running uncontrolled experiments on public roads with no federal pre-market safety approval requirement. NHTSA cannot require a manufacturer to demonstrate AV safety before deployment — it can only respond after harm occurs. The Cruise incident (2023), the Uber fatal crash (Elaine Herzberg, Tempe AZ, 2018), and NHTSA's finding of 400+ Tesla Autopilot crashes in 10 months (2022 Standing General Order data) all represent harm that occurred in the absence of pre-market safety requirements. The fundamental argument: when a product operates in shared public space and can cause death, the public has a compelling interest in requiring safety demonstration before deployment — not merely after. | 88 | 84% | High |
| SAE Level 3-5 autonomy systems require federal standards because vehicles cross state lines and inconsistent state regulation creates a compliance burden that prevents national deployment. If AV systems can legally operate in Arizona (permissive framework) but cannot legally operate in California (more restrictive framework) or New York (limited authorization), manufacturers cannot deploy nationally without navigating 50 different regulatory regimes. The result is either fragmentation (robotaxis only in permissive states) or paralysis (manufacturers wait for federal clarity before investing in national deployment). Federal standards — even permissive ones — create the interstate uniformity that enables nationwide commercial deployment, which is the precondition for AVs realizing their promised safety, mobility, and efficiency benefits at scale. | 85 | 80% | High |
| Insurance liability frameworks are broken for AV crashes and cannot be repaired without federal regulatory action. Current auto insurance is personal liability insurance — it assumes a human driver made a decision. When software makes the driving decision, the relevant liability is product liability (manufacturer of a defective product), not personal auto liability. But: (1) courts are split on whether AV software is a "product" subject to strict liability or a "service" subject to negligence; (2) multi-party causation (sensor manufacturer, software developer, mapping provider, fleet operator) makes standard tort apportionment unworkable; (3) insurers cannot price AV risk without statistical history that doesn't yet exist. The result: AV manufacturers currently operate in a liability gray zone that neither provides adequate victim compensation nor creates efficient safety incentives. Federal legislation that assigns clear liability to AV manufacturers for software-caused accidents is the prerequisite for a functioning AV insurance market. | 82 | 77% | High |
| The documented safety potential of fully autonomous vehicles — if deployed with validated safety — justifies the regulatory investment. The Insurance Institute for Highway Safety (IIHS) estimates that automated crash prevention technologies could prevent up to 40% of fatal crashes if deployed at scale. Human error is implicated in 94% of serious crashes (NHTSA). If Level 4 AV systems can perform within their ODD with a crash rate materially below the human driving baseline (as Waymo's data suggests for its operational domain), the public health benefit of accelerating validated deployment is enormous: 38,000+ annual U.S. traffic deaths are the cost of the status quo. The regulatory question is not whether to allow AVs, but how to create standards that accelerate deployment of safe systems while blocking deployment of unsafe ones. | 80 | 75% | High |
| Clear federal standards paradoxically accelerate AV development by providing regulatory certainty for investment. AV manufacturers cite regulatory uncertainty (unclear standards, unpredictable state-level approval processes, undefined liability allocation) as a primary constraint on investment, alongside the technical challenges. The SELF DRIVE Act, despite failing to pass, generated more regulatory clarity discussions between NHTSA and manufacturers than the preceding decade of voluntary guidelines. Federal standards reduce the compliance risk premium embedded in AV investment decisions, potentially expanding the investment pool and accelerating deployment timelines — the opposite of what opponents who frame regulation as innovation-hostile typically predict. | 76 | 72% | Medium |
| Total Pro (Σ Argument × Linkage): | 320 | ||
❌ Top Scoring Reasons to Disagree | Argument Score | Linkage Score | Impact |
|---|---|---|---|
| Premature federal regulation could lock in 2024 technology standards that become obstacles to 2030 improvements. AV technology is developing rapidly: sensor costs (LiDAR) have fallen from $75,000 to under $500 in a decade; camera-only approaches (Tesla) and sensor-fusion approaches (Waymo) are developing in parallel with different safety profiles. A federal standard written for 2024 technology architectures (requiring LiDAR, specifying minimum sensor specifications, defining "safe" by 2024 accident datasets) could lock in choices that are suboptimal for 2030 technology. The FDA drug approval model — which requires proving safety and efficacy before market entry — takes 10+ years and costs billions, making it unsuitable for software that can be updated continuously. AV-specific regulation needs to be written for software, not hardware, and performance-based rather than design-prescriptive. | 82 | 77% | High |
| NHTSA's existing FMVSS authority is adequate for hardware safety standards, and the software autonomy gap is narrower than regulation advocates claim. AVs still need brakes, lights, crashworthiness structures, and tires — all of which are covered by existing FMVSS. The specific gap is pre-market approval for software-driven dynamic driving tasks, which is indeed a gap — but it is narrow relative to the comprehensive regulatory framework advocates propose. NHTSA's voluntary AV guidance frameworks (AV 1.0, 2.0, 3.0, 4.0) have created de facto standards without binding regulation, and the AV industry largely follows them. The Cruise incident was a corporate culture and incident reporting failure, not evidence of inadequate standards — no regulation would have prevented the crash itself, and existing reporting requirements (violated by Cruise) are the appropriate accountability mechanism. | 76 | 71% | Medium |
| State-level experimentation has generated better regulatory information than a premature federal standard would. California (most restrictive, requiring disengagement reporting and extensive testing documentation), Arizona (most permissive, enabling Waymo and other commercial operations), and Florida (intermediate, focusing on liability allocation) represent a natural experiment in regulatory approaches. The variation is informative: Waymo's operational success in Arizona provides evidence about what permissive-but-monitored deployment produces; California's data collection requirements have generated the most comprehensive AV performance dataset in the world (California DMV Autonomous Vehicle Testing Reports). A federal standard imposed on this experimental variation would flatten the data collection while providing less useful regulatory information. | 73 | 68% | Medium |
| Tort liability for AV accidents already provides stronger safety incentives than regulation under current law. Product liability creates strict liability for manufacturing defects in products — if AV systems are classified as products (the likely outcome after initial litigation), manufacturers face strict liability for crashes attributable to system failures, regardless of whether federal safety standards were met. The financial exposure from a wave of AV product liability suits dwarfs regulatory fines. This is a stronger safety incentive than regulation because it is automatic, unlimited in principle, and applies to each individual crash — unlike regulation, which operates through periodic enforcement actions against the manufacturer. | 70 | 64% | Medium |
| Regulating an immature technology creates competitive disadvantage against Chinese AV developers operating under different constraints. Baidu Apollo, WeRide, and other Chinese AV companies are deploying at scale in Chinese cities under a regulatory framework that prioritizes national industrial policy over precautionary safety assessment. If U.S. federal standards impose pre-market approval requirements that add 3-5 years and $500M+ to each AV system deployment, U.S. developers will fall further behind Chinese competitors in accumulated deployment experience, which is a primary determinant of system improvement. The national security and economic competitiveness implications of Chinese AV dominance (AVs as military dual-use technology, AV platform as data collection infrastructure) argue for regulatory approaches that accelerate U.S. deployment, not constrain it. | 68 | 63% | High |
| Total Con (Σ Argument × Linkage): | 254 | ||
Net Belief Score: +66 (320 Pro − 254 Con) — Moderately Supported; public-road safety, interstate uniformity, and liability-framework arguments give regulation a meaningful edge, though technology-lock-in and competitive-disadvantage concerns are substantial enough to keep this in the moderate range.
⚖ Evidence Ledger
Evidence Type: T1=Peer-reviewed/Official, T2=Expert/Institutional, T3=Journalism/Surveys, T4=Opinion/Anecdote
| Supporting Evidence | Quality | Type | Weakening Evidence | Quality | Type |
|---|---|---|---|---|---|
| NHTSA Standing General Order 2021-01 Crash Data (2022) Source: National Highway Traffic Safety Administration (T2/Official). Finding: NHTSA's mandatory crash reporting order (SGO 2021-01) required AV manufacturers to report crashes involving Level 2+ automation within 24 hours (with casualties) to 30 days. First 10-month report documented 392 crashes involving Tesla Autopilot (Level 2), 130 Waymo (Level 4), 90 GM Cruise (Level 4), and others. Important context: the 392 Tesla Autopilot crashes occurred at a different severity profile than the 130 Waymo crashes — Tesla crashes are more frequent but largely lower-severity (following-distance violations, lane departures); Waymo crashes are less frequent but often involve novel scenarios. The raw numbers cannot be compared without normalization for miles driven and ODD — but the data establishes that AV-involved crashes are occurring and being monitored. |
86% | T2 | RAND Corporation, "Measuring Automated Vehicle Safety: Forging a Framework" (2016) Source: RAND Corporation, independent policy research (T1). Finding: The RAND study calculated that demonstrating AV safety at a statistically meaningful level above the human driving baseline would require between 100 million and 275 billion test miles — far more than is operationally feasible before widespread deployment. This study is the strongest single piece of evidence against FDA-style pre-market approval for AVs: the statistical testing requirement to prove safety by conventional trial is simply impractical. It doesn't argue against regulation — it argues against pre-market approval based on testing as the safety mechanism, suggesting performance monitoring during deployment is necessary regardless. |
85% | T1 |
| NTSB, "Collision Between Vehicle Controlled by Developmental Automated Driving System and Pedestrian" (Tempe, AZ, 2018) Source: National Transportation Safety Board investigation report (T2/Official). Finding: The NTSB found that the Uber AV that struck and killed Elaine Herzberg had: (1) a safety driver who was distracted for 5.7 of the 6 seconds before impact; (2) an automated emergency braking system that had been deliberately disabled by Uber to reduce "erratic" braking behavior during testing; (3) no regulatory requirement that Uber demonstrate a minimum safety culture or testing protocol before deploying on public roads. The crash was preventable by multiple independent safety mechanisms that were absent or disabled. The NTSB specifically recommended that NHTSA establish a framework requiring AV developers to demonstrate safety management systems before operating on public roads. |
88% | T2 | California DMV Autonomous Vehicle Disengagement Reports (2019–2024) Source: California Department of Motor Vehicles (T2/Official). Finding: Annual California DMV reporting of AV disengagements (instances where human driver resumes control) shows dramatic improvement in Waymo performance over 2019–2024: disengagement rate per 1,000 miles fell from 0.076 (2019) to 0.010 (2023) — a 7-fold improvement. This improvement occurred under California's permissive-but-monitored approach (no pre-market approval, strong data reporting). The evidence supports the argument that deployment-with-monitoring is a viable safety framework — the technology improves through deployment, and monitoring catches the failures. This is a cautionary argument against prescriptive pre-market standards that might have prevented the early deployments that generated the improvement data. |
82% | T2 |
| Insurance Institute for Highway Safety (IIHS), "Potential Benefits of Crash Avoidance Technologies" (2020) Source: Insurance Institute for Highway Safety, independent nonprofit (T1/Research). Finding: Automated crash prevention technologies — forward collision warning, automatic emergency braking, lane departure warning — already demonstrably reduce crash rates in vehicles where they are standard equipment. IIHS estimates that if these Level 1-2 technologies were on all vehicles, 40% of fatal crashes could potentially be prevented. This is evidence of the safety upside of automated driving technology when properly validated, establishing the public health justification for regulatory investment to ensure Level 4 systems meet validated safety thresholds before public deployment. |
84% | T1 | Waymo Safety Report, "Waymo's Safety Methodologies and Safety Readiness Determinations" (2020, updated 2023) Source: Waymo LLC (T2/Industry — note: source is interested party). Finding: Waymo's public safety report documents its testing methodology, simulation approach, and operational safety record. Waymo has logged 20M+ fully autonomous commercial miles (no human backup driver) in Phoenix and San Francisco. Its documented crash rate in commercial operations is below the comparison human driving baseline. The report argues that Waymo's internal safety determination process — not federal pre-market approval — is the appropriate mechanism for validating AV safety, because the company has greater system-specific knowledge than any regulator. Acknowledging source bias: Waymo has a direct financial interest in opposing pre-market approval requirements; the technical arguments for performance-based monitoring over pre-market approval are legitimate, but the advocacy conclusion (industry self-certification is adequate) cannot be accepted without independent verification. |
70% | T2 |
| GAO, "Vehicle Safety: NHTSA Has Taken Steps to Prepare for Automated Vehicles but More Could Be Done" (GAO-21-412, 2021) Source: U.S. Government Accountability Office (T2/Official). Finding: GAO found that NHTSA lacks statutory authority for pre-market safety approval of AV systems, has only post-market recall authority, and has not used its rulemaking authority to establish minimum safety standards for AVs despite multiple voluntary guidance documents. GAO specifically recommended that Congress clarify NHTSA's authority and provide resources for AV oversight. The finding is significant because it comes from the non-partisan Congressional oversight body — not from AV advocates or industry — confirming the regulatory gap that is the central premise of the pro-federal-framework argument. |
87% | T2 | McKinsey Global Institute, "Autonomous Vehicles: The Next Frontier of Mobility" (2024) Source: McKinsey Global Institute, consulting/research (T2). Finding: Global AV market projected at $300–400B by 2035. U.S. companies (Waymo, Tesla, Cruise) currently lead global AV technology but Chinese competitors (Baidu Apollo, WeRide, Didi) are rapidly closing the gap in deployment experience, particularly in Level 4 robotaxi operations. The competitive trajectory analysis suggests that without regulatory frameworks that enable U.S. commercial deployment at scale, accumulated deployment experience (which is the primary technical development lever for AV improvement) will favor Chinese developers. The national security and economic implications are cited as arguments for permissive regulation that accelerates deployment. |
74% | T2 |
🎯 Best Objective Criteria
| Criterion | How to Measure | Validity % | Reliability % | Importance |
|---|---|---|---|---|
| AV crash rate vs. human baseline | Crashes per million vehicle miles traveled for Level 4 AV systems in commercial operation, compared to human driving baseline in comparable conditions (urban, mixed traffic). Measured via NHTSA SGO data and California DMV disengagement reports. Effective federal framework success: AV crash rate at or below human baseline within 5 years of framework enactment for commercially deployed systems. | 85% | 82% | High |
| Speed of commercial deployment | Number of jurisdictions with active Level 4 commercial AV operations (robotaxi, freight, delivery) and total commercial AV fleet size, tracked annually. A framework that enables safe deployment should produce measurable expansion of commercial operations; a framework that is excessively prescriptive should show stagnation or contraction. Comparison: U.S. commercial AV fleet vs. Chinese commercial AV fleet growth rates. | 78% | 80% | High |
| Liability resolution time for AV crashes | Average time from AV crash occurrence to victim compensation (settlement or judgment), compared to conventional vehicle crash resolution times. Clear federal liability allocation should reduce resolution time; continued ambiguity should increase it. Track via insurance industry claims data and court records. | 72% | 70% | Medium |
| Safety management system adoption | Percentage of AV developers with NHTSA-certified safety management systems (analogous to aviation SMS requirements) within 3 years of a framework requiring them. SMS adoption rate is a proxy for safety culture institutionalization — the underlying factor in both the Uber and Cruise incidents. | 80% | 75% | High |
| Incident reporting compliance | Rate of compliance with mandatory crash reporting (NHTSA SGO 2021-01 successor) and accuracy of reported information. The Cruise incident involved inaccurate reporting to regulators — tracking and enforcing reporting accuracy is foundational to the entire regulatory regime. Track: reports filed vs. crashes identified through other means (police reports, news coverage); investigations initiated for reporting violations. | 82% | 85% | High |
🔎 Falsifiability Test
| Conditions That Would Confirm the Belief | Conditions That Would Disconfirm the Belief |
|---|---|
| Federal AV framework enactment produces measurable improvement in: crash rate per mile for deployed AV systems (declining toward and below human baseline), expansion of commercial AV deployment in interstate commerce and urban robotaxi operations, and reduction in AV crash-to-compensation time — all within 7 years of enactment, without a significant reduction in AV technology investment that would indicate the regulation was excessively prescriptive. | Federal AV framework enactment is followed by significant AV investment flowing to China, EU, or other less-regulated jurisdictions — measured by a reduction in U.S. share of global AV patent filings, venture investment in AV companies, and commercial AV fleet deployments — confirming that the regulatory costs exceeded the benefits of certainty. |
| The absence of a federal framework continues to produce incidents like the Cruise 2023 crash and the Uber 2018 crash — incidents that were preventable by safety management requirements that no current regulation mandates — confirming that the voluntary guidance framework is inadequate and that mandatory requirements would have changed behavior. | AV companies operating under voluntary guidelines demonstrate safety records that are equivalent to or better than the projected outcomes of mandatory frameworks — confirmed by NHTSA SGO data showing declining crash rates per mile for voluntarily-guided systems — suggesting that mandatory standards provide no marginal safety benefit over voluntary compliance by market-incentivized companies. |
| Clear federal liability allocation (manufacturer liable for software-caused crashes) enables the formation of a functioning AV insurance market with actuarially-priced products, within 3 years of liability clarification — confirming that regulatory uncertainty, not technology immaturity, was the primary obstacle to insurance market development. | Federal liability allocation for AV crashes has no measurable effect on insurance market development — AV-specific insurance products remain unavailable or unaffordable — suggesting that the obstacle to AV insurance is technology immaturity (insufficient statistical data) rather than legal uncertainty, and that liability clarification is premature. |
📊 Testable Predictions
Beliefs that make no testable predictions are not usefully evaluable. Each prediction below specifies what would confirm or disconfirm the belief within a defined timeframe and using a verifiable method.
| Prediction | Timeframe | Verification Method |
|---|---|---|
| At least one additional major AV-involved fatality or serious safety incident attributable to inadequate safety management requirements will occur in the U.S. before federal AV safety management standards are enacted — confirming that the current voluntary guidance framework is inadequate to prevent preventable incidents. | Before federal framework enactment (ongoing) | NHTSA SGO 2021-01 crash reports; NTSB investigation reports; state DMV incident reports. The prediction is conservative — given current deployment scale and the documented Cruise/Uber precedent, additional incidents are historically expected under the voluntary framework. |
| Within 3 years of federal AV framework enactment (if enacted), the number of U.S. jurisdictions with active commercial Level 4 robotaxi operations will at least double from 2024 levels — confirming that regulatory clarity accelerates deployment rather than constraining it, because the framework eliminates state-level compliance uncertainty. | 3 years post-enactment | NHTSA and state DMV commercial AV permit data; annual AV industry deployment surveys (AV industry association, Brookings Institution Metropolitan AV tracking); comparison with pre-framework deployment rate. |
| Chinese AV companies (Baidu Apollo, WeRide, Didi) will achieve greater accumulated commercial deployment mileage (Level 4, driverless) than U.S. companies (Waymo, Amazon Zoox, Motional) by 2028 — unless U.S. federal standards are enacted that enable national commercial deployment — confirming the competitive deployment gap that is the strongest national-security argument for permissive federal standards. | By 2028 | Annual commercial AV mileage data from each company's public safety reports, regulatory filings, and independent AV deployment trackers (BloombergNEF AV tracker, Waymo/Baidu comparative analyses); peer-reviewed academic AV market research. |
| A clear federal AV liability allocation law will enable AV-specific commercial insurance products to reach market within 18 months — with actuarially differentiated pricing by AV system, ODD, and operator safety record — confirming that legal uncertainty, not technology immaturity, is the primary obstacle to AV insurance market development. | 18 months post-liability legislation | Insurance industry filings for AV-specific commercial liability products; NAIC (National Association of Insurance Commissioners) AV insurance market survey; comparison with current insurance market availability for conventional commercial fleets |
⚖ Conflict Resolution Framework
9a. Core Values Conflict
| Federal AV Framework Supporters | Federal AV Framework Opponents (Industry / Anti-Prescriptive) |
|---|---|
| Advertised values: Public safety on shared infrastructure, government accountability for technology deployed in public space, victim compensation for AV-caused harm, equitable access to AV mobility benefits (not just wealthy early adopters). | Advertised values: Innovation-enabling regulation, technology neutrality (not locking in 2024 standards), private sector safety investment efficiency, national competitiveness in AV technology, faster deployment of safety-improving technology. |
| Actual values (in tension): Safety advocates have a genuine interest in AV safety but sometimes advocate for regulatory approaches (extensive pre-market testing requirements) that the RAND study demonstrates are statistically impossible — making their preferred policy framework unrealizable in practice. The actual value in tension: preventing individual vivid harms (Elaine Herzberg, the Cruise victim) may require accepting more diffuse harms (delaying deployment of systems that could prevent 40% of traffic fatalities). The calculus of vivid vs. statistical deaths is uncomfortable but central to the policy question. | Actual values (in tension): AV industry advocates for permissive regulation — but the industry's actual safety record (Cruise incident reporting violations, Tesla Autopilot crash rate, Uber safety culture failure) demonstrates that voluntary safety management produces insufficient safety investment. The industry's actual value (maximizing deployment speed and minimizing regulatory compliance cost) conflicts with the safety claims it advances. The advocacy for permissive regulation is primarily financial self-interest dressed in innovation language — but the innovation language is not entirely without merit. |
9b. Incentives Analysis
| Interests of Federal AV Framework Supporters | Interests of Federal AV Framework Opponents |
|---|---|
| Pedestrian safety advocates. Disability rights organizations (AVs could dramatically expand mobility for non-drivers). Insurance companies (clear liability allocation enables market development). Labor organizations concerned about AV impact on trucking and delivery employment. Urban planners seeking predictable AV integration into street design. Members of Congress from states with AV incidents. NHTSA career staff seeking expanded authority. Families of AV crash victims (Elaine Herzberg family, Cruise victim). Academic safety researchers. Trial lawyers (private right of action for AV crashes). | Waymo, Tesla, Uber ATG successor, Motional, Amazon Zoox (commercial deployment would be constrained by pre-market requirements). AV component manufacturers (Mobileye, Nvidia, Luminar). State governments with established AV industries (Arizona — Waymo; Michigan — Detroit auto industry; California — Silicon Valley) who want to preserve their state regulatory flexibility. Free-market think tanks (Cato, Reason Foundation). Some transportation economists who see regulatory costs as exceeding benefits at current deployment scale. |
9c. Common Ground and Compromise
| Shared Premises | Synthesis / Compromise Positions |
|---|---|
| All major stakeholders agree: the current regulatory vacuum is unsatisfactory and some federal framework is needed. Even the AV industry supported the SELF DRIVE Act in 2017 because it would have provided national uniformity. All stakeholders agree: the framework should be performance-based rather than design-prescriptive (specifying crash rate targets, not technology architectures). All stakeholders agree: mandatory crash reporting is appropriate and the Cruise incident reporting violations were indefensible. All stakeholders agree: federal preemption of state law is necessary for commercial AV deployment at national scale. The genuine dispute is about: whether pre-market approval is required before deployment, how liability is allocated for software-caused crashes, and how rapidly federal standards should restrict state-level regulatory flexibility. | Safety Management System (SMS) requirement: Modeled on aviation SMS requirements — not pre-market approval, but mandatory documentation and certification of safety management processes, incident reporting systems, and safety culture assessments before commercial deployment. This addresses the Cruise and Uber failure modes (safety culture, incident reporting violations) without the testing impracticability problem of pre-market approval. Mandatory incident reporting with criminal penalties for falsification: The Cruise incident reporting violation was the most outrageous element of the 2023 incident. Criminal liability for filing false AV incident reports with NHTSA is the simplest, least controversial legislative fix with universal support. Federal liability allocation: AV manufacturer bears product liability for crashes where software is the proximate cause; fleet operators bear liability where operational decisions deviate from approved ODD. This assigns liability to the party with the greatest ability to prevent the harm. |
9d. ISE Conflict Resolution (Dispute Types)
| Dispute Type | The Specific Disagreement | Evidence or Argument That Would Move Both Sides |
|---|---|---|
| Empirical | Do AV systems actually have lower crash rates than human driving in comparable conditions? Industry says yes (Waymo data); independent researchers say the comparison is confounded by ODD selectivity (AVs operate in favorable conditions by design). | Independent analysis of NHTSA SGO data normalized for miles driven, ODD characteristics, time of day, weather, and traffic density — producing apples-to-apples crash rate comparison. If Waymo's crash rate is materially below human baseline in comparable conditions, this validates the safety case for accelerated deployment. If the crash rate advantage disappears after controlling for ODD selectivity, it confirms that deployment-with-monitoring is the only path to real-world safety data. |
| Empirical | Would a federal Safety Management System (SMS) requirement have prevented the Cruise incident or the Uber crash? Industry argues no — both incidents involved human factor failures (distracted operator, disabled safety system) that regulatory requirements cannot prevent in real-time. Safety advocates argue yes — SMS certification creates organizational accountability structures that change the cost-benefit calculation for disabling safety systems. | Aviation safety data provides the strongest natural experiment: the FAA's SMS requirements for airlines (phased in 2015–2018) are associated with measurable improvements in safety culture indicators and incident rates in the subsequent 5 years. If the aviation SMS pattern holds for AVs (organizational safety culture changes reduce incident rates even when technology is unchanged), the SMS argument is empirically supported. |
| Values | Who bears the risk of AV deployment: the public (involuntarily exposed to AV crashes on public roads) or the companies deploying AV systems (voluntarily accepting financial and reputational risk of crashes)? | This is a genuine values question about risk distribution between innovators and the public. The consent framing is key: Elaine Herzberg did not consent to be a test subject for Uber's AV system. The public road is shared infrastructure, not a private testing ground. The values question is whether commercial innovation that uses shared public infrastructure can externalize its testing risk onto non-consenting members of the public — which is what current voluntary deployment does. This framing shifts most moderate observers toward supporting minimum safety management requirements before public deployment. |
| Definitional | What counts as "federal regulatory framework"? Industry supports a framework that provides national uniformity and preempts state law. Safety advocates want a framework that includes pre-market approval or equivalent safety demonstration requirements. These are substantially different policies called by the same name — and the ambiguity has allowed both sides to claim support for "federal regulation" while meaning opposite things. | An operationally precise definition: a federal AV framework must specify (a) what is required before commercial deployment (SMS certification is minimum; pre-market safety approval is maximum); (b) what mandatory reporting is required during deployment; (c) how liability for crashes is allocated; and (d) whether state laws are preempted or whether states may establish stronger requirements. Requiring proponents to specify their position on each element would end the definitional ambiguity and make it clear that "support for federal regulation" covers a spectrum from SELF DRIVE Act-style permissiveness to FAA-level oversight. |
💡 Foundational Assumptions
| Required to Accept the Belief (Federal AV Framework Is Warranted) | Required to Reject the Belief (Current Voluntary Framework Is Adequate) |
|---|---|
| When technology is deployed in shared public space and can cause death, the public has a compelling interest in requiring safety demonstration before deployment — not merely compensation after harm. This is the foundational premise of every pre-market safety standard in U.S. law (FDA drug approval, FAA aircraft certification, CPSC product safety). AV deployment on public roads is categorically closer to aviation (shared public infrastructure, potential for mass casualties) than to most consumer products. | The pace of AV technology improvement is so rapid that regulatory requirements lock in standards that become obsolete before they can be updated, and the economic and social benefits of accelerated AV deployment (reducing the 38,000 annual traffic deaths) outweigh the marginal safety risks of deploying before formal regulatory validation — particularly when market incentives (tort liability, reputational risk) provide equivalent safety incentives. |
| Voluntary industry safety guidelines produce insufficient safety investment because the financial cost of compliance (delayed deployment, additional testing) exceeds the expected cost of accidents under current tort law — meaning companies rationally under-invest in safety relative to socially optimal levels. This is the classic market failure requiring regulatory correction. | AV companies are rationally investing in safety because: (1) product liability exposure for AV crashes is potentially enormous; (2) reputational damage from AV incidents (GM Cruise program shutdown, Uber ATG closure) demonstrates market accountability; (3) regulatory capture risk — if AV companies lobby for weak standards, they also legitimize the regulatory framework that could eventually impose strong standards. Companies prefer to invest in voluntary safety to avoid regulatory intervention, and this incentive produces adequate safety investment. |
| A performance-based federal framework (specifying crash rate targets and SMS requirements, not technology architecture) can be written to be technology-neutral and adaptive — not locking in 2024 standards for 2030 technology — while still providing the safety floor and liability clarity that the current vacuum lacks. | Any federal regulatory framework will, in practice, be captured by incumbent AV companies and structured to create compliance barriers that favor current market leaders over newer entrants — producing not safety improvement but competitive protection. The history of transportation regulation (taxi medallion systems, airline deregulation) suggests that regulatory frameworks initially designed for safety are systematically converted to competitive protection mechanisms. |
📈 Cost-Benefit Analysis
| Factor | Benefits | Costs / Risks | Likelihood | Impact |
|---|---|---|---|---|
| Traffic safety (human life) | If Level 4 AVs with validated safety profiles replace human driving, potential to prevent 10,000–15,000 annual fatalities (40% of 38,000) at full deployment scale. This is the dominant benefit calculation — statistical lives saved at scale dwarf other considerations. | If federal framework delays deployment of safe systems by 3-5 years due to compliance requirements, the opportunity cost is 30,000–75,000 preventable traffic deaths during the delay period. This is the strongest argument against prescriptive pre-market standards. | 75% meaningful safety improvement at full deployment scale (uncertain on timeline) | Very High |
| Mobility access expansion | Approximately 600,000 Americans with disabilities lose driving capability annually; AV deployment expands mobility access for elderly and disabled non-drivers, reducing healthcare costs and isolation. Potentially the highest-value equity benefit of AV technology. | AV mobility access may initially be available only in wealthy urban areas, widening the mobility gap between urban and rural, wealthy and poor populations before it narrows it. | 65% meaningful mobility expansion within 10 years of deployment scale | High |
| Regulatory compliance costs | Federal uniformity reduces AV company compliance costs for state-by-state approval — estimated $50–200M per jurisdiction currently, generating fragmentation that slows deployment. | SMS certification and reporting requirements add $20–100M per AV developer per system, depending on framework specificity. These costs are passed through to deployment timelines and consumer pricing. | Net compliance cost reduction for large AV companies; net cost increase for smaller AV entrants | Medium |
| Labor displacement (trucking, delivery, taxi) | AV deployment in freight and delivery generates efficiency gains ($70B+ annually in trucking cost reductions projected by ATA) that reduce consumer goods prices and improve supply chain resilience. | 3.5M professional truck drivers and 600,000+ ride-share drivers face employment displacement. Displacement is gradual but concentrated in specific demographic groups (male, rural, lower-education) with limited retraining options. Federal framework should include transition support provisions. | 80% significant employment displacement in long-haul trucking within 15 years of commercial AV freight deployment | High |
| National security / competitive position | Maintaining U.S. AV technology leadership has dual-use military and intelligence applications (autonomous logistics, surveillance, urban operations). Federal investment in AV safety standards can be structured to also advance U.S. competitive position. | Risk: regulatory framework that delays U.S. commercial deployment while Chinese competitors accumulate deployment experience creates a technology gap with long-term strategic implications beyond the commercial AV market. | 40% significant national security impact from AV technology leadership gap | High |
Short vs. Long-Term Impacts
In the short term, a federal AV framework primarily changes the compliance environment for AV developers — creating clarity, enabling insurance market development, and establishing the incident reporting culture that the Cruise incident demonstrated is absent. In the medium term, if the framework is performance-based and appropriately calibrated, it should accelerate national commercial deployment by eliminating state-by-state compliance fragmentation. In the long term, the framework's most important effect may be on liability allocation: clear manufacturer liability for software-caused crashes creates the financial incentive for sustained safety investment that current tort law ambiguity does not.
Best Compromise Solution
Near-term achievable: (1) Criminal penalties for false AV incident reporting — universal support, addresses the most outrageous aspect of the Cruise incident, no legitimate opposition; (2) Mandatory Safety Management System certification before commercial Level 4 deployment — modeled on aviation SMS, performance-based, not design-prescriptive. Medium-term: (3) Federal liability allocation — manufacturer product liability for software-caused crashes within ODD; (4) Federal preemption of conflicting state laws with state reporting and data collection authority preserved. Deliberately defer: pre-market safety approval (impractical given RAND statistical requirements) and mandatory technology specifications (locks in 2024 standards).
🚫 Primary Obstacles to Resolution
These are the barriers that prevent each side from engaging honestly with the strongest version of the opposing argument. They are not the same as the arguments themselves.
| Obstacles for Federal AV Framework Supporters | Obstacles for Permissive Deployment Advocates (Industry) |
|---|---|
| The statistical deaths problem: Safety advocates focus on vivid, identifiable harms (Elaine Herzberg, the Cruise victim) while being unable to give equal moral weight to the statistical deaths that AV deployment delay causes. If safe Level 4 systems deployed 3 years sooner would prevent 30,000 traffic deaths, every year of prescriptive regulatory delay that is not safety-necessary is a policy choice to accept 10,000 preventable deaths. Regulatory advocates who cannot engage with this calculation — or who dismiss it as speculative — are not engaging honestly with the full safety calculus. The policy question is always: compared to what? The alternative to AV deployment is human driving, which kills 38,000 Americans per year. | Incident reporting violations as credibility destroyer: GM Cruise's decision to file inaccurate incident reports with California DMV after the October 2023 crash destroyed the industry's primary argument for self-regulation: that companies have sufficient self-interest in safety to regulate their own behavior. The incident revealed that the financial incentive to avoid operational suspension was strong enough to override both ethical obligations and regulatory compliance. Industry advocates who continue to oppose mandatory safety management requirements after the Cruise incident cannot credibly argue that voluntary compliance is adequate — they have to acknowledge that the current incentive structure failed catastrophically in a real-world case. |
| Pre-market approval as a red herring: Some safety advocates have attached to the FDA drug approval model as a template for AV regulation, despite the RAND study's finding that statistical validation of AV safety requires 100M–275B test miles — making FDA-style pre-market approval operationally impossible. By advocating for an impractical safety standard (pre-market approval based on testing), these advocates provide industry with an easy target to defeat while avoiding the practical achievable safety requirements (SMS certification, liability allocation) that would actually change behavior. The fixation on an unachievable standard prevents progress on achievable ones. | Confusing technology neutrality with regulatory minimalism: AV industry advocates use "technology neutrality" (don't specify sensor types, technology architectures) as code for "no requirements." A performance-based standard that requires a demonstrated crash rate below a specified threshold is technology-neutral — it doesn't specify how to achieve the performance requirement. Industry opposes performance standards as well as design standards, which reveals that "technology neutrality" advocacy is not principled neutrality but opposition to any mandatory requirement. This makes it impossible for industry to negotiate in good faith because any specific compromise proposal can be characterized as "not technology neutral." |
| Constituency misalignment: The strongest advocates for AV safety regulation are pedestrian advocates, disability rights organizations, and families of AV crash victims — constituencies with limited political power relative to the AV industry, which includes major technology companies and the automotive industry. The political asymmetry means that safety regulations are harder to achieve than their merits warrant, because the beneficiaries of regulation (diffuse public safety) are less organized than the costs-bearers (concentrated industry compliance costs). Safety advocates underestimate how much this asymmetry shapes what is legislatively achievable, leading to advocacy for comprehensive frameworks that have no path to passage. | National security argument is selectively applied: AV industry invokes the China competitive threat when opposing safety regulations, but the same competitive threat argument would support substantial government investment in AV safety research and development — which would accelerate U.S. deployment more than regulatory minimalism. If the competitive concern were genuine, the industry position would be "reduce regulation AND increase government investment in AV safety R&D." The position is actually "reduce regulation and do not increase government investment" — revealing that the competitive argument is used selectively to oppose regulation rather than genuinely to advance national competitiveness. |
🧠 Biases
| Biases Affecting AV Framework Supporters | Biases Affecting Permissive Deployment Advocates |
|---|---|
| Omission bias: Human deaths caused by human driving (38,000 per year) are accepted as the baseline; AV deaths (a small fraction of that number) receive disproportionate attention because they are attributable to an identifiable agent (the AV company's software decision). The psychological tendency to weight harmful actions more than equivalent harmful omissions (not deploying safe AV systems) produces systematic bias against AV deployment even when the deployment calculus favors it. | Overconfidence bias: AV developers systematically overestimate the robustness of their systems to edge cases and novel scenarios. This overconfidence is demonstrated by the repeated pattern of AV incidents that companies initially characterized as within their tested performance envelope, only to discover that edge cases (jaywalking pedestrians, construction zones, unusual weather) were inadequately represented in testing. The commercial deployment of insufficiently validated systems is partly driven by genuine technical overconfidence, not merely financial pressure. |
| Availability heuristic: The Uber 2018 crash and Cruise 2023 incident are vivid and memorable; the statistical deaths prevented by deployed AV systems operating safely (Waymo's safety record in Phoenix) are invisible. Safety advocates disproportionately focus on AV failures while underweighting AV successes, because failures generate news coverage and investigations while prevented crashes leave no trace. | Regulatory capture anticipation: AV companies that publicly oppose regulation sometimes privately welcome specific regulatory frameworks that raise compliance barriers against new entrants. The largest AV companies (Waymo, which has spent $10B+ on development) can absorb SMS certification costs that smaller startups cannot. Opposition to federal regulation may therefore partly reflect opposition to the specific proposed frameworks rather than opposition to all frameworks — but the public advocacy position is generalized anti-regulation. |
| Authority framing: The comparison to aviation safety (FAA pre-certification of aircraft) makes AV regulation seem more established than it is. Aviation safety is achieved through a combination of design certification, manufacturer SMS, air traffic control, weather monitoring, and pilot licensing — a multi-layer system that took 70 years to build. Mapping this to AV regulation in a 5-year legislative window is an authority transfer that overstates the applicability of aviation models to road vehicle software. | Deployment pressure: AV companies have invested billions in AV development and face investor pressure to generate revenue from commercial deployment. This financial pressure creates incentives to deploy at the frontier of technical readiness rather than waiting for full validation — which is what the Uber ATG deployment in Tempe (inadequate safety culture, disabled emergency braking) reflected. Companies cannot be relied upon to self-certify the adequacy of their own safety management when the financial pressure to deploy is strong and the probability of regulatory consequence is low. |
🎞️ Media Resources
| Type | Supporting Federal AV Framework | Opposing / Cautionary | Rating |
|---|---|---|---|
| Book | How Cars Became Safe — Joseph Stromberg (longform journalism). The history of automobile safety regulation (seat belts, airbags, crashworthiness standards) shows that manufacturer safety investment was inadequate until mandated — the relevant historical precedent for the AV safety framework debate. | Autonomy: The Quest to Build the Driverless Car — Lawrence Burns (2018). Former GM executive's account of the AV development journey. Provides the industry perspective on why prescriptive regulation would have prevented the technical development that made AV safety possible in the first place. | 8/10 |
| Report | NTSB Uber AV Crash Investigation Report (2019). Available at NTSB.gov. The definitive investigation of the Elaine Herzberg fatality. Specific findings about safety culture failures, disabled emergency braking, and regulatory gaps are the foundation for the safety management requirement argument. Essential primary source reading. | RAND Corporation, "Measuring Automated Vehicle Safety" (2016). The strongest single analytical document for why pre-market safety approval is impractical. Forces safety advocates to confront the statistical testing requirements for their preferred regulatory model. Available at rand.org. | 9/10 |
| Report | GAO-21-412, "Vehicle Safety: NHTSA Has Taken Steps to Prepare for Automated Vehicles but More Could Be Done" (2021). Non-partisan Congressional oversight finding that confirms the regulatory gap. The most credible source for the specific legislative authority gap claim. Available at GAO.gov. | Waymo Safety Report (2020, updated 2023). The most comprehensive public documentation of a Level 4 AV safety methodology. Even skeptics of self-certification should read it to understand what sophisticated voluntary safety management looks like — and what minimum mandatory standards would need to require to add value beyond it. | 8/10 |
| Article | Alexis Madrigal, "Inside Waymo's Secret World for Training Self-Driving Cars" (The Atlantic, 2017). Accessible explanation of AV simulation-based testing — why billions of test miles are run in simulation, and why this doesn't substitute for real-world edge case validation. Useful for understanding the testing methodology debate. | Matthew Wansley, "The Regulation of Emerging Risks" (Vanderbilt Law Review, 2021). Academic analysis of how regulatory frameworks applied too early to immature technologies systematically produce suboptimal outcomes — the most rigorous academic case for the "wait for technology maturity" position on AV regulation. | 8/10 |
⚖ Legal Framework
| Laws and Frameworks Supporting Federal AV Regulation | Laws and Constraints Complicating Federal AV Regulation |
|---|---|
| National Traffic and Motor Vehicle Safety Act (49 U.S.C. §§ 30101–30170): The foundational federal vehicle safety statute, establishing NHTSA's authority to set Federal Motor Vehicle Safety Standards and issue recalls. Provides existing legal authority for hardware-focused AV regulation (sensors, braking systems, passenger protection). Does not provide authority for pre-market software approval — the legislative gap that AV-specific legislation would need to fill. Recall authority requires demonstration of a safety defect post-deployment, making it adequate only for post-incident response. | Interstate Commerce Clause and NHTSA preemption: Federal vehicle safety standards preempt conflicting state laws under existing statute. However, NHTSA's current voluntary AV guidelines explicitly do not preempt state law — leaving states free to regulate AV deployment conditions. Mandatory federal AV standards would preempt state law under existing NHTSA preemption doctrine, but voluntary guidelines do not. The transition from voluntary to mandatory standards therefore involves a simultaneous change in preemption status — which is a significant political obstacle because states that have established AV regulatory frameworks (California, Arizona) are reluctant to cede authority to federal standards they did not participate in designing. |
| SELF DRIVE Act (H.R. 3388, passed House 2017) / AV START Act (S. 1885, 2017): The closest Congress has come to federal AV legislation. The SELF DRIVE Act passed the House unanimously in 2017 but died in the Senate. Key provisions included: NHTSA authority to exempt AV systems from FMVSS requirements, mandatory Safety Assessment Letters (voluntary submission of safety demonstration documentation), preemption of state AV performance standards, and establishment of an AV working group. The bill's failure is instructive: it died because Senate Commerce Committee members objected to preemption provisions and inadequate labor displacement provisions, not because of fundamental disagreement about whether federal AV legislation was needed. | Products Liability Law (state common law): AV product liability litigation is governed by state common law — there is no federal product liability statute. Courts in different states apply different standards (strict liability vs. negligence for design defects, different standards for software-as-product vs. service). Without federal liability allocation, AV crash victims must litigate in state courts under varying standards, and AV manufacturers face different liability exposures in different states. Federal AV legislation that includes liability allocation provisions would preempt state products liability standards — a significant jurisdictional change that the plaintiff's bar (and the Senate Democrats aligned with it) has historically opposed. |
| Federal Aviation Administration Authorization Act (49 U.S.C. § 40101 et seq.) — Aviation SMS as model: The FAA's Safety Management System requirements for commercial aviation (14 CFR Parts 5, 119, 121) provide the regulatory template most frequently cited for AV safety management. Aviation SMS requires airlines to document safety risk management processes, maintain safety assurance systems, and conduct regular safety performance reviews. Adapted to AVs: a pre-deployment SMS certification requirement would require AV developers to demonstrate adequate safety management culture and processes before commercial deployment — addressing the Uber and Cruise failure modes without the testing impracticability of pre-market approval. | Administrative Procedure Act (5 U.S.C. §§ 551–559) — Rulemaking timeline: Even if Congress grants NHTSA authority to establish mandatory AV safety standards, APA notice-and-comment rulemaking requirements mean that final rules will take 3–7 years to promulgate after authority is granted. AV technology will have changed substantially by the time any rule based on 2024 legislative authority takes effect. This practical constraint supports performance-based standards (which can be written at a technology-neutral level of generality that remains valid as technology changes) over design-prescriptive standards (which specify technology requirements that may be obsolete before the rule takes effect). |
| FAV FORWARD Act (2021) and subsequent AV proposals: Multiple AV-specific bills have been introduced since SELF DRIVE failed, all addressing similar provisions: NHTSA authority expansion, SMS requirements, mandatory reporting, and liability allocation. The repeated introduction of similar bills demonstrates bipartisan consensus on the need for legislation, while repeated failure to pass demonstrates the specific obstacles (preemption, labor, liability forum). The accumulated legislative record of failed AV bills provides a detailed map of what consensus exists and what specific provisions have blocked passage. | First Amendment (compelled speech) — Incident reporting: Mandatory incident reporting requirements that compel companies to provide detailed information about crashes and near-misses implicate First Amendment compelled speech doctrine. While the government's interest in safety information generally survives First Amendment scrutiny, the scope of mandatory reporting (real-time data access, behavioral logs, pre-crash data recording) in some legislative proposals pushes toward ongoing mandatory data transmission that courts might characterize as government-compelled surveillance of commercial activity rather than post-incident disclosure. |
🔗 General to Specific Belief Mapping
| Relationship | Belief | Connection |
|---|---|---|
| Upstream (broader) | When private companies deploy technology in shared public space that can cause death, the public has a compelling interest in requiring safety validation before deployment — regardless of the technology's commercial promise. | The foundational claim about public safety vs. private commercial innovation that underlies not just AV regulation but pharmaceutical approval, nuclear power oversight, and airline safety certification. |
| Upstream (broader) | AI systems with consequential real-world effects should be subject to mandatory safety assessment before deployment. (See belief_ai-regulation.html) | AV software is an instance of high-stakes AI deployment — the most consequential real-world AI system currently operating at commercial scale, in terms of direct physical harm potential. The AV regulatory question is a specific version of the general AI safety governance question. |
| Downstream (specific) | Congress should pass the SELF DRIVE Act or equivalent AV legislation granting NHTSA authority to require Safety Assessment Letters and safety management system certification before Level 4 commercial deployment. | The specific legislative proposal that has achieved the greatest bipartisan support — more specific than the general claim that federal AV regulation is warranted. |
| Downstream (specific) | Filing false or misleading incident reports with NHTSA about AV crashes should be a federal criminal offense with personal liability for corporate officers. | The most immediately achievable and least controversial specific component of an AV regulatory framework — addressing the Cruise 2023 incident reporting violation directly. |
| Related lateral | The United States should invest in public transit infrastructure and reduce car dependency. (See belief_invest-transit.html) | AV regulation and transit investment address the same underlying question (how should Americans move around?) from different angles. AV deployment may complement transit (first/last mile) or compete with it (private AV ownership reduces transit demand). The interaction effects should be considered in any comprehensive transportation policy analysis. |
💡 Similar Beliefs (Magnitude Spectrum)
| Positivity | Magnitude | Belief |
|---|---|---|
| +85% | 78% | The United States should require pre-market safety certification for all Level 4+ AV systems before any public road deployment, modeled on FAA aircraft certification with NHTSA as the certifying authority. No commercial deployment without demonstrated statistical safety validation. (Maximalist — FAA-level pre-market approval, despite RAND statistical impracticability.) |
| +58% | 68% | The United States should establish a federal regulatory framework for autonomous vehicles including: mandatory Safety Management System certification before commercial deployment, mandatory incident reporting with criminal penalties for falsification, federal liability allocation for software-caused crashes, and preemption of conflicting state laws. (THIS BELIEF — performance-based framework, SMS-centered, without pre-market approval.) |
| +40% | 55% | Congress should grant NHTSA authority to require voluntary Safety Assessment Letters from AV developers and establish national data standards for AV crash reporting, without mandatory SMS requirements or federal preemption of state deployment rules. (Moderate — expand reporting and data, preserve state authority, no mandatory safety management requirements.) |
| +25% | 45% | Federal AV regulation should be limited to data standards and incident reporting coordination, with deployment authority remaining at the state level and voluntary industry safety guidelines serving as the primary safety mechanism. The current NHTSA voluntary guidance framework is adequate for the current deployment scale. (Minimal federal role — preserve state experimentation, voluntary safety framework.) |
| -5% | 50% | Any mandatory federal AV regulatory framework — beyond existing FMVSS and post-deployment recall authority — would lock in premature standards that constrain beneficial innovation and should be deferred until AV technology matures to a point where performance standards can be written without technology prescription risk. (Anti-prescriptive — defer all mandatory standards until technology matures.) |
No comments:
Post a Comment