General Motors’ self-driving unit Cruise is facing investigations by the Department of Justice and Securities and Exchange Commission after one of its vehicles dragged a pedestrian across a San Francisco street in November 2023. An internal report released this week reveals multiple failures in technology, safety protocols, leadership and disclosure around the alarming event.
The Pedestrian Dragging that Shocked the Nation
On November 12, 2023 at approximately 10:15PM, a Cruise autonomous vehicle made a right turn in the Mission district of San Francisco when it struck a man crossing the street [1]. Video footage shows the man trapped underneath the vehicle as it continues down the street, dragging him over 100 feet before coming to a stop [2].
The pedestrian was hospitalized with serious injuries. While an initial police report stated that the individual may have been intoxicated when entering the crosswalk against the light, an ongoing investigation has uncovered deeper issues with Cruise’s technology and response [3].
| Key Details of the Cruise Pedestrian Dragging Incident |
|-|
| Date | November 12, 2023 |
| Time | Approx. 10:15PM |
| Location | Mission District, San Francisco |
| Cruise Vehicle | Modified Chevrolet Bolt |
| Victim Details | Adult male pedestrian |
| Victim Injuries | Hospitalized with serious injuries |
The event shocked the nation and immediately halted Cruise’s driverless taxi service in San Francisco, which had launched at the beginning of 2023. Experts called into question the safety of autonomous vehicles (AVs), especially those operating without human safety drivers [4].
Intense scrutiny prompted federal investigations while Cruise pursued an internal review to uncover what went wrong. Details from these probes paint an alarming picture of technological and organizational failures.
Department of Justice Investigates Safety Reporting Lapses
The internal Cruise report summarizes underlying factors that likely led to the collision and post-impact dragging incident [5]. However, the Department of Justice (DOJ) investigation focuses on Cruise’s lapses in safety procedures and disclosure around autonomous vehicles.
The DOJ probe centers on two key questions:
-
Did Cruise properly disclose vehicle safety defects? Engineering logs revealed Cruise was aware of issues related to post-collision behavior of its AVs weeks prior to the pedestrian incident. The DOJ aims to determine if these qualify as safety defects that should have been reported.
-
Did Cruise file accurate safety assessment reports? All companies testing autonomous vehicles on public roads must submit regular Federal Automated Vehicles Safety Reports. Cruise’s reports stated its vehicles would come to a gradual stop after most collisions – conflicting with real-world behavior.
Cruise’s failure to address known AV issues prior to the pedestrian collision, paired with inaccurate safety filings, could constitute negligence under federal law. The DoJ investigation likely aims to uncover if criminal charges are warranted [6].
Securities Exchange Commission Questions Leadership and Disclosures
In tandem with the DoJ investigation, the Securities Exchange Commission (SEC) launched a probe into the leadership and disclosures around Cruise’s autonomous vehicle operations [7].
The SEC appears concerned that Cruise misled investors by downplaying risks associated with testing robotaxis on public roads without human safety drivers. Specifically, they are focused on statements made by Cruise CEO Kyle Vogt regarding the readiness of technology and safety of operations.
In the months leading up to launch of driverless taxi service, Vogt stated:
“We feel very confident that the vehicle is able to handle the complexities of driving in San Francisco.”
Following the pedestrian dragging incident, an internal report revealed Cruise’s executive team was aware of potential scenarios the vehicles could not safely handle. The SEC aims to determine if Vogt knowingly mischaracterized technology capabilities despite awareness of limitations [8].
This probe serves as a warning signal to AV startups and automakers that the SEC expects full transparency around progress and safety. As emerging technology intersects with public markets, leadership teams must carefully qualify known risks even when the business incentivizes hype.
Internal Report Uncovers Multiple Failures That Led to Incident
Beyond federal investigations, Cruise pursued an internal review to uncover root causes of the collision and subsequent dragging incident. The report exposes technological shortcomings as well as lapses in training and protocols across the organization [9].
AV Technology Could Not Reliably Categorize Vulnerable Road Users
Cruise’s autonomous vehicles use cameras, radar and lidar to perceive and categorize objects around them. However, the internal report found “systemic deficiencies” in how the software identifies and predicts movements of pedestrians, cyclists and other vulnerable road users [10].
In the moments leading up to collision, video footage shows the victim entered the crosswalk against the light with a bag in hand. He then stopped part way across the opposite lane. Logs show Cruise’s perception system struggled to accurately categorize the object, flipping between labels of “debris” and “pedestrian” in the seconds before impact.
The AV technology’s inability to reliably identify the presence of a pedestrian appears to have been a key factor leading to the tragic collision.
Post-Collision Behavior Exposed Flaws in Operational Design Domain
Operational Design Domain (ODD) refers to the specific conditions under which an autonomous vehicle system is designed to function properly and safely. Cruise’s ODD allowed vehicles to operate at night, in light rain, and make unprotected turns up to 30 mph [11].
The internal report concluded that Cruise’s ODD was inadequately defined to prevent unsafe post-collision behavior. Although vehicles had been engineered to safely come to a stop after most minor impacts, the system did not account for potential entrapments of smaller vulnerable road users underneath the vehicle.
Cruise essentially designed AVs capable of normal driving tasks but failed to consider rare but dangerous edge cases. The pedestrian dragging incident exposed flawed assumptions that contributed to unsafe operations outside the true functional ODD.
Lacking Protocols and Training for Safety Critical Scenarios
In addition to limitations in core AV technology, Cruise lacked adequate protocols and training for safety critical scenarios among software engineers, fleet response teams and corporate leadership.
The report cited ineffective coordination between remote monitoring agents and in-field safety drivers leading to delays in response. Additionally, Cruise’s executive safety committee lacked clearly defined procedures for incidents involving serious injuries or fatalities.
Internal communication and leadership breakdowns exacerbated the already dangerous situation. Cruise vows to implement safety focused culture change through updated protocols, training and leadership policies [12].
Cruise Halts Robotaxi Service While Major Reforms Underway
The convergence of technological failures and organizational unpreparedness proved catastrophic, shaking public faith in the safety of autonomous vehicles. Cruise’s internal report clearly states that the company and technology was not ready for fully driverless commercial taxi service.
In response to federal probes and its own findings, Cruise announced major reform efforts alongside continuance of the earlier robotaxi program suspension [13]:
Technological Improvements
- Upgrading perception systems for more reliable recognition of pedestrians and cyclists
- Expanding scenario testing to improve safety of post-collision behaviors
- Enhancing simulation models to better reveal edge case deficiencies
Operational Changes
- Implementing new safety driver protocols and training programs
- Revising ODD limitations to restricthigher risk driving scenarios
- Instating executive incident response protocols for major collisions
Corporate and Culture Updates
- Safety focused transformation of internal policies and training
- Leadership accountability for transparent disclosures
- Cross-functional collaboration to improve safety culture
Cruise states it will resume driverless taxi operations only after rigorous internal and external testing proves readiness. Given the fallout of rushing immature technology into commercial service, years may pass before Cruises AVs return sans human safety drivers.
The industry waits anxiously to see if Cruise can truly deliver substantive reforms after such an alarming safety wake up call. But optimism remains that autonomous vehicles will eventually transform transportation once rigorous diligence ushers the technology safely forward.
References
- https://www.nytimes.com/2024/01/25/technology/cruise-crash-report-san-francisco.html
- https://www.theverge.com/2024/1/25/24050791/cruise-pedestrian-dragging-video-driverless-report
- https://www.wired.com/story/robot-car-crash-investigation-cruise-disclose-key-information/
- https://finance.yahoo.com/news/gm-cruise-owns-robotaxi-mess-200023339.html
- https://www.nytimes.com/2024/01/25/technology/cruise-crash-report-san-francisco.html
- https://breakingthenews.net/Article/GM’s-Cruise-confirms-probe-by-DoJ-over-dragging-incident/61336403
- https://finance.yahoo.com/news/gm-cruise-robotaxi-unit-faces-200315879.html
- https://www.marketscreener.com/news/latest/THE-SUBJECT-AUTONOMOUS-VEHICLE-S-MAINTENANCE-RECORDS-DID-NOT-SHO–45815967/
- https://techcrunch.com/2024/01/25/cruise-reveals-doj-sec-probes-as-it-releases-internal-report-on-pedestrian-crash/
- https://www.theverge.com/2024/1/25/24050791/cruise-pedestrian-dragging-video-driverless-report
- https://www.marketscreener.com/news/latest/ROOT-CAUSE-OF-CRUISE-VEHICLE-S-POST-COLLISION-MOVEMENT-AFTER-IN–45816017/
- https://www.cnbc.com/2024/01/25/gm-cruise-probe-finds-poor-leadership-at-center-of-incident-response.html
- https://finance.yahoo.com/news/gm-cruise-robotaxi-unit-faces-200315879.html
To err is human, but AI does it too. Whilst factual data is used in the production of these articles, the content is written entirely by AI. Double check any facts you intend to rely on with another source.
