Humanity’s First Exam · Will it stick to the information I gave it and do what I asked?
You asked for 500 words and got more than twice that
You asked for a briefing of about 500 words. Grok Bot wrote 1,417, with a 19-row table and a long list of caveats. Claude Fable 5.2 wrote 1,270.
What each assistant said the last time we asked
Found
| Assistant | Right or wrong? |
|---|---|
| ChatGPT dots | Got it right |
| Grok Bot | Got it wrong |
| Instinct | Got it right |
| Muse | Got it right |
| Claude Fable 5.2 | Got it wrong |
All the times we’ve asked this
- Claude Fable 5.2 asked. Claude Fable 5.2: wrong.
- Found in our review. ChatGPT dots: right; Grok Bot: wrong; Instinct: right; Muse: right.
Each time we ask again, the result and its date are added here.
What to check yourself. If length matters, say so, and check the reply against it.
What the assistant was asked
The exact words it was given:
Use facts and sources available on or before 18 October 2026 unless the task specifies another period or version. Use only evidence you can actually access. Do not buy, book, contact anyone, schedule work or change accounts. Return the answer here, linking the evidence for material claims. Prepare an evidence briefing comparing two November 14, 2023 road-safety publications: NTSB report HIR-23-09 on the North Las Vegas crash and IIHS’s research release on vehicle front-end design and pedestrian fatalities. Describe the intervention each institution argues for, the mechanism and evidence behind it, and the strengths and limits of comparing them. Explain whether the proposals address the same problem in the same way and what an editor should say about their regulatory status. Aim for 500 words, with a hard maximum of 700 words excluding references, an evidence table, and a brief search log. Cite the primary sources you actually used with section, table, or page locators.
ChatGPT dots Got it right
See what ChatGPT dots said
These publications support complementary interventions, not competing explanations of one crash. NTSB targets excessive speed and repeat offending; IIHS targets the injury consequences of vehicle-front geometry. Both seek fewer deaths, but operate at different points in the chain of events. Evidence at a glance PublicationEvidence and populationLocatorNTSB HIR-23-09, dated November 14, 2023January 29, 2022 multivehicle crash: a Dodge reached 103 mph on a 35-mph road and ran a red light; nine vehicle occupants died. Event recorders, surveillance, toxicology and driver records informed the investigation.Report, pp. 1–5, 13–20IIHS, November 14, 202317,897 single-vehicle/single-pedestrian crashes, plus detailed reconstruction of 121 crashes involving teenage or adult pedestrians.Release, paragraphs beginning “To examine the connection” and “To better understand”NTSB: reduce dangerous speed before impact. Its probable-cause finding identifies excessive speed and disobedience of traffic controls, with drug impairment and failures to deter repeat offending contributing. Intelligent speed assistance (ISA) identifies local speed limits using mapping/GPS and sign recognition. Advisory systems warn; active versions intervene in speed. NTSB recommends standard ISA that at minimum warns, not universally non-overridable governors. It separately proposes research and guidance for speed-limiting interlock pilots for repeat offenders. HIR-23-09, §§2.5.2, 3.2 and 4.1, pp. 24, 37–38. The report combines case reconstruction with earlier ISA studies showing reduced speeding. It says active limitation might have mitigated this crash, not that warning-only equipment certainly would have prevented it. Effects depend on engagement, acceptance and implementation. §2.5.2, pp. 25–29. IIHS: make the striking vehicle less injurious. IIHS urges manufacturers to lower hood leading edges and use more sloping fronts. Compared with low, sloped fronts, hoods over 40 inches were associated with approximately 44–45% greater pedestrian fatality risk; medium-height blunt fronts with 26% greater risk. These are relative comparisons among crashes, not increases in crash frequency. Analyses excluded pedestrian-AEB vehicles and controlled for factors including posted speed limits, age and sex. Detailed reconstructions linked tall, blunt fronts with torso/hip injuries and forward projection rather than rolling onto the hood. Release, methods, results and reconstruction paragraphs. What comparison can establish. The strengths differ: NTSB provides a detailed account of a particular disaster and a technology-policy argument; IIHS combines a larger observational comparison with evidence about injury mechanisms. Neither is a randomized head-to-head test of the proposed interventions. Posted speed limits are not measured impact speeds, and adjustment does not eliminate every possible confounder. An editor should therefore present speed management and safer geometry as potentially complementary, not substitute one for the other or multiply their reported effects. Comparative cost-effectiveness would require comparable populations, actual implementation, exposure measures and injury outcomes. The publications do not supply that evaluation. Regulatory status and dating. Write “NTSB recommended that NHTSA require ISA,” not “NTSB mandated ISA.” Its separate manufacturer request sought installation, while IIHS encouraged redesign rather than announcing a binding standard. These publications do not establish subsequent regulatory adoption. NTSB announcement, recipient-specific recommendations; IIHS release, manufacturer recommendations. A dating nuance: November 14 is the NTSB report’s stated date and board-action date; the same-day announcement said the full report would be posted several weeks later. It should not be described as necessarily available in full that day. NTSB announcement, closing paragraphs. Brief search log: Located and read HIR-23-09, the contemporaneous NTSB announcement and IIHS’s dated release. Used the report’s printed page numbers. Excluded later pedestrian studies and later regulatory responses from this comparison.
Grok Bot Got it wrong
Both publications argue for changing the vehicle, but at different points in a crash. NTSB wants intelligent speed assistance (ISA, technology that detects the speed limit and warns the driver or holds the car to it) in all new vehicles, to cut speeding. IIHS wants lower, sloped front ends, so that a pedestrian who is hit is less likely to die. Neither publication is binding. NTSB: intelligent speed assistance The crash. On 29 January 2022, a Dodge Challenger ran a light that had been red for at least 29 seconds. It reached a recorded 103 mph in a 35 mph zone and struck a minivan. Nine people died: both people in the Dodge and all seven in the minivan (N1, abstract; §1.1 pp.1, 4). Probable cause: "the Dodge driver's excessive speed and failure to obey traffic control devices." Contributing factors: the driver's cocaine and PCP impairment, and Nevada's "failure to deter the driver's speeding recidivism" (§3.2 p.37). The intervention. Recommendation H-23-14 asks the National Highway Traffic Safety Administration (NHTSA) to "require as standard equipment in all new vehicles intelligent speed assistance systems that, at a minimum, warn the driver when the vehicle exceeds the speed limit." H-23-20 asks 17 automakers to fit the same systems. Other recommendations cover: teaching drivers about ISA; ISA interlocks for repeat speeders; state programmes for repeat offenders; adding ISA to NHTSA's car ratings (H-17-24, reiterated) (§4.1–4.2 pp.38–39). Mechanism and evidence. How it works: ISA reads the speed limit from map data and cameras. Passive systems warn the driver, while active systems resist or cap speed (§2.5.2 p.24). What the evidence shows: it's about driving behaviour, not crash counts. In one study, the share of distance driven over the limit fell from 28% to 9% with warnings, and from 26% to 5% with active ISA. NTSB infers that crashes "likely would also be reduced." The benefits lasted only while the systems were switched on (p.25). What NTSB says about this crash: NTSB doubted that a warning would have slowed this impaired driver. It concludes only that an active system "may have mitigated the severity" of the crash (p.25; Finding 7, p.36). My observation: the minimum NTSB recommends, a warning, is weaker than the active system it thinks might have helped here. IIHS: front-end design The intervention. IIHS wants automakers to lower the front edge of the hood and slope the front end. Lead author Wen Hu said there is "no functional benefit to these massive, blocky fronts" (I1, ¶8, ¶16). Evidence. The study covered 17,897 police-reported crashes in seven states, each involving one vehicle and one pedestrian. Front ends of 2,958 vehicle models were measured from photos. Vehicles with pedestrian automatic emergency braking were excluded, and the analysis controlled for speed limit and the pedestrian's age and sex (I1 ¶6; I2). Compared with low, sloped fronts, the risk of the pedestrian dying rose by: about 45% for sloped hoods over 40 inches high; 44% for blunt hoods over 40 inches; 26% for blunt fronts of medium height; 25% for flat hoods (I1 ¶7, ¶9–10). Reconstructions of 121 crashes suggest why: tall, blunt fronts strike the torso and hip and throw pedestrians forward (I1 ¶11–14; I3). Comparing them Strengths. Both came out the same day, and both use vehicle design to make up for human error across the US fleet. They're also directly linked: NTSB asks IIHS to study car marketing that promotes risky driving (H-23-21; N1 p.39). Limits. Different evidence bases: NTSB starts from one crash and supports ISA with studies of driving behaviour, not estimates of crash reduction. IIHS's study is observational, so it shows an association, not cause and effect. It also adjusts for posted speed limits rather than actual impact speed (my reading). Different problems: ISA targets speeding, which endangers everyone on the road. Front-end design reduces harm only after a pedestrian has been hit, and no pedestrian was involved in the Las Vegas crash. Together? The two measures could complement each other, but neither publication tests that (my inference). Regulatory status: what an editor should say NTSB recommends; NHTSA regulates. NHTSA describes itself as "the federal agency that sets and enforces vehicle safety standards" and lists H-23-14 as open (R2). The report notes NHTSA "has not published performance standards for ISA" (p.26). US: as of 15 October 2026, NHTSA's ISA work was research into user acceptance that "is not immediately intended to inform rulemaking" (R1). I found no US requirement for ISA. An older NTSB call for test criteria for pedestrian-friendly vehicle designs (H-18-41) is also still open (R2). EU: Regulation (EU) 2019/2144 requires ISA. Drivers can switch it off, and it must not stop them exceeding the limit. It applied to new vehicle types from 6 July 2022 and to all new registrations from 7 July 2024 (Art. 6(2); Annex II row D8, note B). IIHS is a nonprofit funded by auto insurers, not a regulator (I1, page footer). Evidence table # Claim Source Locator Type 1 Crash on 29 Jan 2022; light red for at least 29 s; 103 mph recorded; 9 killed N1 Abstract; §1.1 pp.2, 4; §1.2.1 p.5 Fact 2 35 mph limit; "almost three times the legal limit" N1 §1.1 p.1; §2.5.2.1 p.25 Fact and quote 3 Probable cause and contributing factors N1 §3.2 p.37; pp.v–vi Quote 4 H-23-14 to 21; H-17-24 reiterated N1 §4.1–4.2 pp.38–39 Quote 5 How ISA works; passive vs active N1 §2.5.2 p.24 Paraphrase 6 Over-limit distance 28%→9% and 26%→5%; crashes "likely would also be reduced" N1 p.25 Figures and quote 7 Only active ISA "may have mitigated the severity" N1 p.25; Finding 7, p.36 Quote 8 No NHTSA performance standards for ISA N1 p.26 Quote 9 17,897 crashes; 2,958 models; seven states; controls used I1; I2 I1 ¶6; I2 abstract Fact 10 +45%, +44%, +26%, +25% fatality risk I1; I2 I1 ¶7, ¶9–10; I2 abstract Figures 11 121 crash reconstructions and the injury mechanism I1; I3 I1 ¶11–14; I3 abstract Paraphrase 12 Hu quote I1 ¶8 Quote 13 IIHS funded by insurers I1 Page footer Fact 14 NHTSA sets standards; H-23-014 and H-18-041 open R2 Intro; table rows Quote and fact 15 ISA research "not immediately intended to inform rulemaking" R1 91 FR 65614–19 Quote 16 No US ISA requirement found R1, R2, N1 — My inference (based on finding no evidence of one) 17 EU ISA rules and dates E1 Art. 6(1)(a), 6(2)(b),(d); Annex II D8 Quote and fact 18 Speed only partly controlled in the IIHS study I1 ¶6 — My reading 19 The two measures could complement each other — — My inference References N1: NTSB, Multivehicle Crash at Signalized Intersection, North Las Vegas, Nevada, January 29, 2022, report HIR-23-09, 14 Nov 2023. Page numbers are the printed ones; body p.1 is PDF p.8. N2: NTSB, NTSB Calls for Technology to Reduce Speeding in All New Cars, news release, 14 Nov 2023. I1: IIHS, Vehicles with higher, more vertical front ends pose greater risk to pedestrians, 14 Nov 2023. Paragraph numbers are my own count, starting from the lede. I2: Hu, Monfort and Cicchino, Journal of Safety Research, 2024, abstract. I3: Monfort, Hu and Mueller, Traffic Injury Prevention, 2024, abstract. R1: NHTSA, Intelligent Speed Assistance user acceptance and effectiveness testing, Federal Register notice 91 FR 65614, 15 Oct 2026. R2: NHTSA, NTSB Open Recommendations to NHTSA, live page. E1: Regulation (EU) 2019/2144, original 2019 text. Search log and caveats Fetch date: I fetched everything on 19 Oct 2026, a day after your cutoff. Every document is dated 15 Oct 2026 or earlier. NTSB report version: the PDF's file data shows it was created on 2 Jan 2024. That's the final report posted after the 14 Nov 2023 board meeting. NTSB news release: it mentions "eight new and one reiterated recommendation to [NHTSA]," but the report shows five new ones to NHTSA, with eight new in total. I followed the report. NTSB's EU dates: the report's own description of the EU rules is less precise, so I used the dates in the regulation itself. IIHS papers: I read only the abstracts. The journal versions are dated 2024, after the release. Not verified: NHTSA's April 2024 response to NTSB was blocked, so I didn't read or cite it. NTSB's own current status for H-23-14 is also unverified; the "open" status comes from NHTSA's page. For the EU regulation, I didn't check later amendments. Length: the body is 682 words, under the 700-word maximum but above the 500-word aim. I checked every quote and figure against the source text.
What it said when we asked it to check its answer
Every assistant got the same follow-up:
Recheck your answer against the original request and the sources you can actually access. Correct any unsupported or outdated claim, fill any missing requirement you can verify, and identify anything still unresolved.
Grok Bot replied:
The briefing holds. I rechecked every NTSB quote and figure against the report text, and each is on the page I cited. The IIHS figures are right too, but I described the comparison groups too loosely, and two details on study design need tightening. None of this changes the conclusions. Confirmed NTSB HIR-23-09, using printed page numbers: The speed limit was 35 mph (p.1), and the light was red for "at least 29 seconds" (p.2). The recorded 103 mph appears on p.4. The probable cause wording, including cocaine, phencyclidine (PCP) and Nevada's "failure to deter," is on p.37. The text of H-23-14, H-23-21 and the reiterated H-17-24 is on pp.38–39. The study figures (28% falling to 9%, and 26% to 5%), "likely would also be reduced," "almost three times" and "may have mitigated the severity" are all on p.25. The last of these is repeated in Finding 7, p.36. "Has not published performance standards for ISA" is on p.26. IIHS release: It gives 17,897 crashes and 2,958 models, excludes vehicles with pedestrian automatic emergency braking, and controls for speed limit and the pedestrian's age and sex. The 45%, 44%, 26% and 25% figures, the 121 reconstructed crashes and the Hu quote are all there. The footer says IIHS is "wholly supported by these auto insurers and insurance associations." Corrections IIHS comparison groups: I summarised all four figures as increases over low, sloped fronts. The release words them differently: 45%: tall, sloped vehicles compared with vehicles of "a similar slope and hood heights of 30 inches or less." 44%: tall, blunt vehicles. The release gives the number but doesn't spell out the comparison group in that sentence. 26%: "Compared with low and sloped vehicles." 25%: flat hoods compared with "more sloping hoods," whatever the height or front-end shape. Where two study details come from: "police-reported" and "seven states" come from the paper's abstract (I2), not the 2023 release. That journal version is dated 2024, after the release. What the 121 reconstructed crashes show: this smaller study split vehicles at 35 inches, not 40. Taller vehicles were more dangerous "mainly because they tended to cause more severe head injuries." Torso and hip injuries were the specific pattern for tall, vertical fronts, which were "more likely to injure pedestrians by throwing them forward." I had blurred those two findings together. Requirement check Covered: each institution's proposed intervention, how it's meant to work, the evidence behind it, the strengths and limits of comparing the two, whether they tackle the same problem in the same way, and their regulatory status. Supporting material: the evidence table, references and search log. Length: the body is 682 words, under the 700-word maximum. The corrections add only a few words. Still unresolved NTSB report version: the PDF was created on 2 January 2024. It's the final report, not a file from the 14 November board meeting. NTSB recommendation count: the news release says "eight new" recommendations went to NHTSA, but the report shows five. I followed the report. US regulatory status: "no US requirement for ISA" rests on finding no evidence of one, plus the 15 October 2026 Federal Register notice. NHTSA's April 2024 reply to NTSB was blocked, and NTSB's own current status for H-23-14 is unverified. EU regulation: I checked only the original 2019 text, not later amendments. My own readings: that the two measures could work together, and that the IIHS study only partly controls for speed, are my inferences. Fetch date: I fetched everything on 19 October 2026, a day after your cutoff. Search log: I didn't fetch anything new. I re-ran text matches on the saved NTSB report, the IIHS release and the IIHS abstract pages, and read the full sentences around each corrected passage.
Instinct Got it right
See what Instinct said
NTSB HIR-23-09 vs IIHS front-end design release, both 14 November 2023 Identification - The NTSB item is correct: Highway Investigation Report HIR-23-09, "Multivehicle Crash at Signalized Intersection, North Las Vegas, Nevada, January 29, 2022". The cover shows November 14, 2023. - The IIHS item is a news release, not a standalone report. Its underlying paper (Hu, Monfort, Cicchino, Journal of Safety Research) came out in September 2024, after the release. So the release is the only IIHS primary source from that date. What each argues for - NTSB: NHTSA should require intelligent speed assistance (ISA) as standard in all new vehicles, at minimum a warning when the car exceeds the speed limit (H-23-14). Further recommendations cover a public communication plan, repeat-offender tracking, countermeasures for speeding recidivism and pilot ISA interlock programs for repeat speeders (H-23-15 to -18). The earlier read also lists recommendations to states and manufacturers (H-23-19, -20) and one to IIHS on marketing (H-23-21). NTSB also reiterates H-17-24 on ISA incentives such as NCAP. - IIHS: manufacturers should lower hood leading edges and slope grilles and hoods. The release quotes IIHS saying automakers should consider the findings. Mechanism and evidence - NTSB is a single-crash investigation. A Dodge reached 103 mph in a 35 mph zone, ran a red light and killed nine. Probable cause was excessive speed and failure to obey the signal. Cocaine and PCP impairment and Nevada's failure to deter a repeat speeder were contributing factors. NTSB found ISA "may have mitigated the severity". That is a hedged finding, not a measured effect. The mechanism is behavioural: warn or limit speed. - IIHS is statistical plus biomechanical. It analysed 17,897 single-vehicle, single-pedestrian crashes across 2,958 models, excluding vehicles with pedestrian automatic emergency braking and adjusting for speed limit and pedestrian age and sex. Hood edges over 40 inches were about 45% more likely to be fatal than low, sloped fronts. Blunt fronts at 30-40 inches were 26% more likely, and flat hoods added 25%. A 121-crash reconstruction set suggests tall, blunt fronts cause torso injuries by throwing the pedestrian forward. Same problem, same way? - Partly. Both concern deaths linked to vehicles and both direct asks at NHTSA or manufacturers. But NTSB addresses crash occurrence and severity driven by speed, impairment and recidivism. IIHS addresses injury severity once a pedestrian is struck. They are complementary: lower speed should also reduce pedestrian harm, but neither source tests that link. That is my inference. Strengths and limits of comparing - Same date, same sector, same regulator named. But a case investigation supports causation for one event, not an effect size. IIHS gives adjusted associations from observational data. The 45% figure can't be set against anything in the NTSB report. Regulatory status for an editor - Neither document is a regulation. NTSB recommendations are non-binding. IIHS is a research organisation, and its release makes no regulatory request. - Verified by 18 October 2026: NHTSA's April 2024 reply asked that H-23-14, -15, -16, -18 and H-17-24 be classed "Open, Acceptable Response" and H-23-17 "Closed, Acceptable Response". NTSB's online record for H-23-14 (cached 25 June 2026) shows "Open - Initial Response Received". - That page also flags "Is NPRM: Yes". I did not find a published ISA proposed or final rule and could not verify what the flag refers to. - Safe wording: "NTSB recommended that NHTSA require ISA. The recommendation remains open, and no ISA mandate was verified. IIHS's findings are research, not a rule or a recommendation to a regulator." Not verified - The report's section 2.5.2 and section 4 text (the fetch was truncated). The H-23-19 and H-23-20 numbers and statuses come from other pages and were not confirmed in this check. - Whether NHTSA took any ISA rulemaking or NCAP decision. - The current status of recommendations other than the ones named above.
Muse Got it right
See what Muse said
Two publications released on 14 November 2023 locate road risk in vehicles, not only in drivers or roads, but they intervene at different points in a crash. The National Transportation Safety Board's North Las Vegas report (HIR-23-09) concerns a 29 January 2022 crash in which a 2018 Dodge Challenger entered an intersection against a red light at a recorded 103 mph, hit five other vehicles and killed nine people, including seven minivan occupants. The Board found the cause was excessive speed, drug-impaired driving — the driver was impaired by cocaine and PCP — and Nevada's failure to deter his speeding recidivism despite numerous citations. Its headline intervention is intelligent speed assistance (ISA) in all new cars. ISA compares GPS location with a database of posted limits and uses onboard cameras to help ensure legal speeds. Passive systems warn visually, audibly or haptically, leaving slowing to the driver; active systems make exceeding the limit harder or electronically limit speed. The Board asked NHTSA to require ISA that at a minimum warns, and to pursue public education, tracking of repeat offenders, countermeasures and ISA-interlock guidance for them, and NCAP incentives; it asked states to measurably reduce recidivism, and 17 manufacturers to install warning ISA. The mechanism is prevention — redundancy against human error before impact — in a context where speeding-related crashes caused 12,330 deaths in 2021, about one-third of U.S. traffic fatalities. The Insurance Institute for Highway Safety study asks instead: once a pedestrian is hit, how does front-end geometry affect whether they die? Its intervention is design: lower the front end of the hood and angle the grille and hood to create a sloped profile, a change lead author Wen Hu defended by arguing there is no functional benefit to massive, blocky fronts. The evidence is epidemiological. Researchers analysed 17,897 single-vehicle, single-pedestrian crashes, derived measurements for 2,958 models from photographs via vehicle identification numbers, excluded vehicles with pedestrian automatic emergency braking, and controlled for speed limit and the pedestrian's age and sex. Vehicles with hoods over 40 inches were about 45% more likely to cause a pedestrian fatality than vehicles 30 inches or less with a sloping profile, whatever their nose shape; medium-height (30–40 inch) blunt fronts were 26% more likely to kill than low, sloped vehicles, while medium sloped fronts were not; flat hoods added 25% risk regardless of height or shape. A linked review of 121 detailed crashes supplies the mechanism: vehicles over 35 inches mainly caused more severe head injuries, vertical fronts added more frequent, severe torso and hip injuries, and tall, blunt vehicles tended to throw pedestrians forward rather than roll them onto the hood. Do they address the same problem in the same way? Only partly. Both locate risk in vehicle characteristics that regulators and manufacturers control, against rising deaths — pedestrian deaths up 80% since 2009 in IIHS's framing. But NTSB targets crash causation and speed, generalised from one catastrophic, well-investigated case; IIHS targets injury severity, estimated across thousands of cases. The comparison usefully shows two complementary levers — prevent the high-energy impact, and make the impact that still occurs less lethal. Its limits are clear: different designs (single-case investigation versus observational study), different outcomes (a specific multivehicle crash versus pedestrian fatality risk), an ISA effect asserted from investigative logic rather than estimated in this crash, and IIHS associations that, though controlled, cannot prove a redesign alone would cut deaths by the quoted percentages — notably, sloping did not help the tallest vehicles. For regulatory status, an editor should write recommended and urged, not required or banned. Neither publication is a regulation. NTSB issued recommendations — eight new and one reiterated to NHTSA, plus others to states, manufacturers and IIHS; only NHTSA, or states and manufacturers acting voluntarily, could turn them into requirements. IIHS simply encouraged automakers to reconsider height and shape. As published, neither document created an ISA mandate or a front-end design standard. References NTSB, "NTSB Calls for Technology to Reduce Speeding in All New Cars," news release, 14 Nov 2023 (summarising HIR-23-09 findings and recommendations): https://www.ntsb.gov/news/press-releases/Pages/NR20231114.aspx IIHS, "Vehicles with higher, more vertical front ends pose greater risk to pedestrians," news release, 14 Nov 2023: https://www.iihs.org/news/detail/vehicles-with-higher-more-vertical-front-ends-pose-greater-risk-to-pedestrians Evidence table Institution Publication Intervention argued Mechanism Key evidence (as published) NTSB Nov 14, 2023 release on HIR-23-09, North Las Vegas multivehicle crash Require ISA in all new cars (warning minimum); state recidivism programmes; manufacturer installation ISA prevents the speeding event; redundancy against human error before impact Single crash investigation: 103 mph in a 35 mph zone, red-light violation, 9 killed; driver cocaine/PCP-impaired with repeat speeding history; contextual stat: 12,330 speeding deaths in 2021 IIHS Nov 14, 2023 research release Lower hood leading edge; angled grille/hood (sloped profile) Tall/blunt fronts strike at head/chest height and throw pedestrians forward, raising severe-injury odds 17,897 single-vehicle pedestrian crashes, 2,958 models measured, pedestrian-AEB vehicles excluded, speed limit/age/sex controlled; >40 in hoods ~45% deadlier; medium blunt fronts +26%; flat hoods +25%; 121-crash injury-mechanism review Search log Five web searches: NTSB HIR-23-09 board meeting; IIHS front-end pedestrian release; site-scoped NTSB and IIHS queries; exact-title IIHS phrase; HIR-23-09 PDF. Two pages opened and cited: the NTSB 14 Nov 2023 release and the IIHS 14 Nov 2023 release. The full HIR-23-09 PDF was not located; the NTSB release itself notes the final report would be published weeks later, so NTSB claims here rest on that release's summary of the executive summary, probable cause, findings and recommendations.
Claude Fable 5.2 Got it wrong
**Bottom line.** The two publications share a date and a vehicle-design focus, but they address different problems with different kinds of evidence. Neither is a regulation, and I found no federal rule adopting either as of 18 October 2026. ## NTSB HIR-23-09 - **Intervention.** The Board recommended that NHTSA require intelligent speed assistance (ISA) in all new vehicles that, at minimum, warns the driver over the limit (H-23-14). It asked 17 automakers to fit the same as standard (H-23-20) and sought tracking of repeat speeders and pilot ISA "interlock" programs (H-23-16 to -19). - **Mechanism.** ISA finds the limit from GPS and a speed-limit database plus sign-reading cameras, then either alerts the driver (passive) or resists or limits speed (active). - **Evidence.** One investigated crash: a Dodge Challenger entered a red light at 103 mph on a 35-mph street, and nine people died. The Board adds field trials; in one, distance driven over the limit fell from 28% to 9% with a warning system. - **Limits.** Crash reductions are inferred from speed behaviour, and the effects lasted only while systems were on. The Board was unsure a warning-only system would have slowed this impaired driver and concluded only that an active limiter may have mitigated the crash. It defends the passive minimum as a phased step to build public acceptance. ## IIHS release - **Intervention.** IIHS says manufacturers can lower the hood's leading edge and slope the grille and hood, and encourages automakers to reconsider SUV and pickup front ends. The release requests no rule; the paper says NHTSA can consider such evaluations in its New Car Assessment Program (NCAP) or regulation. - **Evidence.** 17,897 police-reported single-vehicle, single-pedestrian crashes in seven states, with front ends of 2,958 models measured from photos and analysed by logistic regression. - **Findings.** Hoods over 40 inches carried about 45% higher fatality risk than low, sloped fronts; blunt medium-height fronts, 26%; flat hoods, 25%. - **Mechanism.** From a separate 121-crash study: taller fronts caused more severe head injuries, and tall, blunt fronts struck the torso and threw pedestrians forward. - **Limits.** The release says tall vehicles "cause" more deaths, while the abstract says "associated". The outcome is death given a crash, not crash likelihood. Vehicles with pedestrian automatic braking were excluded and the speed control was the posted limit, not impact speed. IIHS is funded by auto insurers. ## Comparing them They do not address the same problem in the same way. - **Stage of harm.** NTSB targets whether an extreme-speed crash happens at all; IIHS targets how badly a pedestrian is hurt once struck. - **Victims.** NTSB's nine deaths were vehicle occupants; IIHS studies pedestrians. - **Evidence type.** NTSB offers a case study plus behavioural literature; IIHS offers a population regression. There is no common effect metric, and NTSB gives no lives-saved estimate. - **Document type.** One is a formal recommendation with tracked agency responses; the other is a news release on research not journal-published until September 2024. - **Common ground.** Both treat vehicle design as the lever and look to NHTSA and automakers. They are complementary rather than competing, since IIHS itself notes speeding has helped fuel the rise in pedestrian deaths. That linkage is my inference; neither document cites the other's remedy. ## Regulatory status: what an editor should say Use "recommended" or "urged", never "required" or "proposed rule". - **ISA.** NTSB cannot regulate. NHTSA's April 2024 reply committed only to two ISA research projects. Its NCAP roadmap keeps ISA in research through 2028. A California passive-ISA mandate was vetoed in September 2024 (trade press; I did not open the veto message). - **Repeat offenders.** NHTSA's April 2026 plan mentions exploring ISA devices for repeat offenders. That is nearer H-23-18 than a new-vehicle mandate. - **Abroad.** The EU requirement, effective July 2024, lets drivers override the system. - **Front ends.** No US rule limits hood height or bluntness. NHTSA's September 2024 proposal (FMVSS 228) is a head-to-hood impact test at about 25 mph. It remained at the proposed-rule stage in the Spring 2025 Unified Agenda. NCAP's pedestrian check was postponed to model year 2027 and is consumer information, not a mandate. - **Unverified.** I could not load NTSB's recommendation database for current classifications, and I checked no Unified Agenda later than Spring 2025. ## Evidence table | Claim | Source and locator | |---|---| | Crash facts: 103 mph, 35-mph limit, nine deaths | HIR-23-09, §1.1 pp. 1–4; Table 1 p. 6 | | Seven speeding citations in five years, four reduced to non-moving violations | HIR-23-09, §2.3 p. 19; Table 2 p. 14 | | ISA definition, passive vs active | HIR-23-09, §2.5.2 p. 24 | | ISA field-trial effects; persistence only while engaged | HIR-23-09, §2.5.2.1 p. 25 | | Passive ISA uncertain for this driver; active "may have mitigated" | HIR-23-09, §2.5.2.1 p. 25; Finding 7 p. 36 | | Phased passive-then-active rationale | HIR-23-09, §2.5.2.2 p. 28 | | Recommendations H-23-14 to H-23-21; H-17-24 reiterated | HIR-23-09, §4.1–4.2 pp. 38–39 | | Probable cause, including Nevada's systemic deficiencies | HIR-23-09, §3.2 p. 37 | | Sample, method, controls, exclusion of automatic-braking vehicles | IIHS release ¶6; paper abstract "Method" | | 45% / 44% / 26% / 25% risk increases | IIHS release ¶¶2, 7, 9, 10; abstract "Results" (45.4, 43.6, 25.6, 25.1) | | Injury mechanism, 121-crash study | IIHS release ¶¶11–15 | | Design advice to automakers | IIHS release ¶¶8, 16; abstract "Practical applications" | | NHTSA response to NTSB | NHTSA letter, 3 Apr 2024, pp. 1–3 | | FMVSS 228 scope and test | 89 FR 76922, §I and §V.E (pp. 76922–76926, 76934) | | FMVSS 228 stage | Unified Agenda, Spring 2025, RIN 2127-AK98 (search-result text only) | | NCAP: ISA in research; pedestrian check delayed | 89 FR (3 Dec 2024) roadmap table; 90 FR 45463 (22 Sep 2025) (search-result text only) | Notes on the table: - The IIHS release has no sections, so paragraph numbers are my count from the headline. - Appendix B of the NTSB report gives page pointers that run about one page later than the printed pagination; cite by section. - A search snippet of an NTSB list shows 1/11/2024 against HIR-23-09. I did not open that file, so the PDF's release date is unconfirmed; the report itself is dated November 14, 2023. ## References 1. NTSB, *Multivehicle Crash at Signalized Intersection, North Las Vegas, Nevada*, HIR-23-09: https://www.ntsb.gov/investigations/AccidentReports/Reports/HIR2309.pdf 2. NTSB investigation page HWY22FH004: https://www.ntsb.gov/investigations/Pages/HWY22FH004.aspx 3. IIHS news release, 14 Nov 2023: https://www.iihs.org/news/detail/vehicles-with-higher-more-vertical-front-ends-pose-greater-risk-to-pedestrians?hl=en-US 4. Hu, Monfort and Cicchino, *Journal of Safety Research*, Sept 2024 (abstract only): https://www.iihs.org/research-areas/bibliography/ref/2294 5. NHTSA response to NTSB, 3 Apr 2024: https://www.nhtsa.gov/sites/nhtsa.gov/files/2024-04/NHTSA-NTSB-Response_April-2024_Intelligent-Speed-Assistance_ISA-Interlock_Speeding_NCAP.pdf 6. NHTSA, proposed rule for FMVSS 228, 89 FR 76922: https://www.govinfo.gov/content/pkg/FR-2024-09-19/pdf/2024-20653.pdf 7. Unified Agenda, Spring 2025, RIN 2127-AK98: https://www.reginfo.gov/public/do/eAgendaViewRule?pubId=202504&RIN=2127-AK98 8. NCAP final decision (driver-assistance systems and roadmap), 3 Dec 2024: https://www.federalregister.gov/documents/2024/12/03/2024-27447/new-car-assessment-program-final-decision-notice-advanced-driver-assistance-systems-and-roadmap 9. NCAP delay notice, 22 Sep 2025: https://www.federalregister.gov/documents/2025/09/22/2025-18285/new-car-assessment-program-ncap-notice-delay-of-program-updates 10. NHTSA, Pathways to Safer Streets release, 21 Apr 2026: https://www.nhtsa.gov/press-releases/pathways-to-safer-streets-launch 11. Repairer Driven News on the SB 961 veto (secondary): https://www.repairerdrivennews.com/2024/10/02/california-gov-vetoes-bill-requiring-speed-assistance-in-vehicles ## Search log - **Read in full:** references 1–5. Reference 6 was read through §VI only. The first NTSB PDF link redirected to an unrelated page; the copy on ntsb.gov worked. - **Search-result text only:** references 7–11. - **Failed:** NTSB's recommendation page for H-23-014 returned no content. - **Searches run:** HIR-23-09 and ISA; the IIHS release title; H-23-14 status; FMVSS 228 status and RIN 2127-AK98; NCAP final decision and delay; the April 2026 DOT plan; the SB 961 veto. - **Not accessed:** the full IIHS journal paper, the companion 121-crash paper, the NTSB docket, and the veto message itself.
What it said when we asked it to check its answer
Every assistant got the same follow-up:
Recheck your answer against the original request and the sources you can actually access. Correct any unsupported or outdated claim, fill any missing requirement you can verify, and identify anything still unresolved.
Claude Fable 5.2 replied:
I rechecked every claim against the sources. The substance of the briefing holds, with three corrections and four sources upgraded from search excerpts to full reads. The corrected briefing is below, followed by what remains unresolved. ## What changed - **Publication date corrected.** NTSB's own 14 November 2023 news release says the full report would be published in several weeks. I previously left this as unconfirmed. - **EU claim softened.** It rests only on NTSB's 2023 forward-looking statement, so it is now attributed that way. - **Citation fixed.** The FMVSS 228 test description pointed to a search excerpt. It now points to the Federal Register text I read. - **Now read in full:** the NTSB news release, the Spring 2025 Unified Agenda entry, the 2025 NCAP delay notice, and NHTSA's April 2026 plan release. ## Corrected briefing (about 640 words) **Bottom line.** The two publications share a date and a vehicle-design focus, but they address different problems with different kinds of evidence. Neither is a regulation, and I found no US federal rule adopting either as of 18 October 2026. **Dating.** On 14 November 2023 NTSB adopted HIR-23-09 and released the executive summary, probable cause, findings and recommendations. The full report came later, so describe it as adopted that day, not published in full. ### NTSB HIR-23-09 - **Intervention.** The Board recommended that NHTSA require intelligent speed assistance (ISA) in all new vehicles that, at minimum, warns the driver over the limit (H-23-14). It asked 17 manufacturers to fit the same as standard (H-23-20). It also sought tracking of repeat speeders and pilot speed-limiting programs for them (H-23-16 to -19). - **Mechanism.** ISA finds the limit from GPS and a speed-limit database plus sign-reading cameras, then alerts the driver (passive) or resists or limits speed (active). - **Evidence.** One investigated crash: a Dodge Challenger ran a red light at 103 mph on a 35-mph street, and nine people died. The Board adds field trials; in one, distance driven over the limit fell from 28% to 9% with a warning system. - **Limits.** Crash reductions are inferred from speed behaviour, and effects lasted only while systems were on. The Board was unsure a warning-only system would have slowed this impaired driver and concluded only that an active limiter may have mitigated the crash. It defends the passive minimum as a phased step to build acceptance. ### IIHS release - **Intervention.** IIHS says manufacturers can lower the hood's leading edge and slope the grille and hood, and encourages automakers to reconsider SUV and pickup front ends. The release requests no rule. The paper's abstract says NHTSA can consider such evaluations in its New Car Assessment Program (NCAP) or regulation. - **Evidence.** 17,897 police-reported single-vehicle, single-pedestrian crashes in seven states, with front ends of 2,958 models measured from photos and analysed by logistic regression. - **Findings.** Hoods over 40 inches carried about 45% higher fatality risk than low, sloped fronts; blunt medium-height fronts, 26%; flat hoods, 25%. - **Mechanism.** From a separate 121-crash study: taller fronts caused more severe head injuries, and tall, blunt fronts struck the torso and threw pedestrians forward. - **Limits.** The release says tall vehicles "cause" more deaths; the abstract says "associated". The outcome is death given a crash, not crash likelihood. Vehicles with pedestrian automatic braking were excluded, and the speed control was the posted limit, not impact speed. IIHS is funded by auto insurers. ### Comparing them They do not address the same problem in the same way. - **Strengths of the comparison.** Both treat vehicle design, not only driver behaviour, as the lever, and both look to NHTSA and automakers. The institutions are also formally linked: NTSB's report asks IIHS to evaluate automaker marketing that emphasises speeding (H-23-21). - **Stage of harm.** NTSB targets whether an extreme-speed crash happens; IIHS targets how badly a pedestrian is hurt once struck. - **Victims.** NTSB's nine deaths were vehicle occupants; IIHS studies pedestrians. - **Evidence type.** NTSB offers a case study plus behavioural literature; IIHS offers a population regression. There is no common effect metric, and NTSB gives no lives-saved estimate. - **Document type.** One is a formal recommendation with tracked agency responses; the other is a news release on research journal-published in September 2024. - **Complementarity.** IIHS notes that speeding has helped fuel the rise in pedestrian deaths. Treating the two remedies as complementary is my inference; neither document cites the other's. ### Regulatory status: what an editor should say Use "recommended" or "urged", never "required" or "proposed rule". - **ISA.** NTSB cannot regulate. NHTSA's April 2024 reply committed only to two ISA research projects. Its NCAP roadmap keeps ISA in research through 2028 (search excerpt only). A California passive-ISA mandate was vetoed in September 2024 (trade press). - **Repeat offenders.** NHTSA's April 2026 plan supports exploring ISA devices for repeat offenders. That is nearer H-23-18 than a new-vehicle mandate. - **Abroad.** NTSB reported in 2023 that EU ISA rules would apply from July 2024 and allow driver override. I did not check EU law. - **Front ends.** No US rule limits hood height or bluntness. NHTSA's September 2024 proposal (FMVSS 228) is a head-to-hood impact test at about 25 mph. The Spring 2025 Unified Agenda lists it at the proposed-rule stage, with further analysis needed before it could be finalised. NCAP's pedestrian check was postponed to model year 2027 and is consumer information, not compliance. ## Evidence table | Claim | Source and locator | |---|---| | Crash facts: 103 mph, 35-mph limit, nine deaths | HIR-23-09, §1.1 pp. 1–4; Table 1 p. 6 | | Driver's citation history; four reduced to non-moving violations | HIR-23-09, §2.3 p. 19; Table 2 p. 14 | | ISA definition, passive vs active | HIR-23-09, §2.5.2 p. 24 | | Field-trial effects; persistence only while engaged | HIR-23-09, §2.5.2.1 p. 25 | | Passive ISA uncertain for this driver; active "may have mitigated" | HIR-23-09, §2.5.2.1 p. 25; Finding 7 p. 36 | | Phased passive-then-active rationale | HIR-23-09, §2.5.2.2 p. 28 | | Recommendations H-23-14 to H-23-21; H-17-24 reiterated | HIR-23-09, §4.1–4.2 pp. 38–39 | | Probable cause, including Nevada's systemic deficiencies | HIR-23-09, §3.2 p. 37 | | Full report to follow the 14 Nov 2023 meeting; 17 manufacturers | NTSB news release, final and recommendation paragraphs | | Sample, method, controls, exclusion of automatic-braking vehicles | IIHS release ¶6; abstract "Method" | | 45% / 44% / 26% / 25% risk increases | IIHS release ¶¶2, 7, 9, 10; abstract "Results" (45.4, 43.6, 25.6, 25.1) | | Injury mechanism, 121-crash study | IIHS release ¶¶11–15 | | Design advice; NCAP or regulation | IIHS release ¶¶8, 16; abstract "Practical applications" | | NHTSA response to NTSB | NHTSA letter, 3 Apr 2024, pp. 1–3 | | FMVSS 228 scope and test | 89 FR 76922, §I pp. 76922–76926; §V.E p. 76934 | | FMVSS 228 stage | Unified Agenda, Spring 2025, RIN 2127-AK98, "Timetable" and "Additional Information" | | Pedestrian check delayed; consumer information, not compliance | 90 FR 45463, "Dates" p. 45463; §III.A p. 45465 | | ISA in NCAP research phase to 2028 | 89 FR 95916 (3 Dec 2024), roadmap table (search excerpt only) | | ISA devices for repeat offenders | NHTSA release, 21 Apr 2026, Pathway 4 | Notes on the table: - The IIHS release has no sections, so paragraph numbers are my count from the headline. - Appendix B of the NTSB report gives page pointers that run about one page later than the printed pagination; cite by section. - The NTSB news release miscounts: it attributes eight new recommendations to NHTSA. The report's §4.1 lists eight in total, five of them to NHTSA. ## References 1. NTSB, HIR-23-09: https://www.ntsb.gov/investigations/AccidentReports/Reports/HIR2309.pdf 2. NTSB news release, 14 Nov 2023: https://www-s.ntsb.gov/news/press-releases/Pages/NR20231114.aspx 3. NTSB investigation page HWY22FH004: https://www.ntsb.gov/investigations/Pages/HWY22FH004.aspx 4. IIHS news release, 14 Nov 2023: https://www.iihs.org/news/detail/vehicles-with-higher-more-vertical-front-ends-pose-greater-risk-to-pedestrians?hl=en-US 5. Hu, Monfort and Cicchino, *Journal of Safety Research*, Sept 2024 (abstract only): https://www.iihs.org/research-areas/bibliography/ref/2294 6. NHTSA response to NTSB, 3 Apr 2024: https://www.nhtsa.gov/sites/nhtsa.gov/files/2024-04/NHTSA-NTSB-Response_April-2024_Intelligent-Speed-Assistance_ISA-Interlock_Speeding_NCAP.pdf 7. NHTSA, proposed FMVSS 228, 89 FR 76922: https://www.govinfo.gov/content/pkg/FR-2024-09-19/pdf/2024-20653.pdf 8. Unified Agenda, Spring 2025, RIN 2127-AK98: https://www.reginfo.gov/public/do/eAgendaViewRule?pubId=202504&RIN=2127-AK98 9. NCAP delay notice, 90 FR 45463: https://www.govinfo.gov/content/pkg/FR-2025-09-22/pdf/2025-18285.pdf 10. NCAP final decision and roadmap, 3 Dec 2024 (excerpt only): https://www.federalregister.gov/documents/2024/12/03/2024-27447/new-car-assessment-program-final-decision-notice-advanced-driver-assistance-systems-and-roadmap 11. NHTSA, Pathways to Safer Streets release, 21 Apr 2026: https://www.nhtsa.gov/press-releases/pathways-to-safer-streets-launch 12. Repairer Driven News on the SB 961 veto (secondary, excerpt only): https://www.repairerdrivennews.com/2024/10/02/california-gov-vetoes-bill-requiring-speed-assistance-in-vehicles ## Search log - **Read in full:** references 1–6, 8, 9 and 11. Reference 7 was read through §VI only. - **Search excerpt only:** references 10 and 12. federalregister.gov blocks automated access. - **Failed:** NTSB's recommendation page for H-23-014 returned no content. Two data.ntsb.gov PDF links redirected to an unrelated page. - **New searches this round:** a later Unified Agenda entry for RIN 2127-AK98, and 2026 news of an FMVSS 228 final rule, withdrawal or delay. Neither returned anything newer than Spring 2025. ## Still unresolved - **Exact release date of the full NTSB report.** The news release confirms it came after 14 November 2023. A search excerpt of an NTSB list suggests 11 January 2024, but I could not open that file. - **Current NTSB classifications of H-23-14 to H-23-21.** I have only what NHTSA requested in April 2024. - **FMVSS 228 after Spring 2025.** I found no final rule or withdrawal, but could not open a later agenda entry. "Still proposed" is my best reading, not a confirmed status at 18 October 2026. - **NHTSA's ISA research.** I did not find results of the two projects, or whether its speed-management guideline was updated in 2025 as planned. - **Status of the EU rule and the California veto message.** Neither was checked against a primary source. - **Full IIHS papers.** I read only the abstract of the main paper and nothing of the 121-crash paper, so confidence intervals, the seven states and the crash years are unverified.
Where this example came from
Our review Can an AI assistant replace your Google search? has the full scores and what the test does not show. The people and households in it are made up.
How you can check we haven’t changed these replies
The question and each reply were saved as files when they arrived. These fingerprints identify those files.
- the question:
d73fe9be149a2570ae9a0777fa0d9d166f5a85e3189e8f282b26cbbbb235bb17 - ChatGPT dots, reply:
c8c9b86e9d1f3341a86ba04f4c5262cfb7f9606908a34de7e6744f58c3e6080e - Grok Bot, reply:
0cd9c3d84d8081cea1f0b6b5ba323263337ea5b4ef4e433274eb0a7df9c8b9e2 - Grok Bot, after the follow-up:
4a9a94e9a5d3349ea89b43e93cf84e1a280231bc82265d2f9c3eacad4b30da2d - Instinct, reply:
017c14f6df733399ff9a8f68fccd20976f1751f2daef076275df16bbc08d010e - Muse, reply:
01e2a8a01aef365f4b1821bf1cdcad3cce62c754d1d10543ad75a954db75e618 - Claude Fable 5.2, reply:
97d5ba3eeecfd73625e78735fdcd0acdf04f3a2e28f31072570d202953c0a9a9 - Claude Fable 5.2, after the follow-up:
16bfb790b149430688db40a46d56d3fe4646c83653b1eb9e3fb8a035e9017f79
Found a mistake in this case? Tell us