At $99 a month, Tesla’s Full Self-Driving (Supervised) subscription gives Canadian owners access to one of the most ambitious driver-assistance packages sold to the public. Tesla says the system can navigate, steer, change lanes, make turns and park under active supervision, while its Canadian website promotes a striking safety message: vehicles using FSD experience far fewer major and minor collisions than the estimated U.S. average.
That claim is facing renewed scrutiny. A Reuters investigation published October 7, 2026, reported that traffic-safety researchers who examined Tesla’s European evidence said the studies did not support broader claims that FSD prevents fatal crashes or saves lives. The dispute matters in Canada because FSD remains a supervised driver-assistance system, not an autonomous chauffeur. The person behind the wheel must continue watching the road, obeying traffic laws and taking over when needed.
What the $99 Subscription Actually Buys
Tesla Canada currently lists FSD (Supervised) at $99 per month. Keeping the subscription active for a full year therefore costs $1,188 before tax. Tesla says eligible vehicles generally need its FSD Computer 3.0 or newer, although some owners may require hardware upgrades that are not included in the subscription price. The company also says the plan can be cancelled at any time, with access continuing until the end of the current billing period rather than generating a prorated refund.
The software offers considerably more than conventional cruise control. Tesla says FSD can follow a navigation route, steer, change lanes, negotiate intersections, enter and exit highways, make turns and assist with parking. But the company repeatedly states that those abilities require active driver supervision and do not make the vehicle autonomous. That distinction is crucial. The name “Full Self-Driving” can sound like a transfer of responsibility, while Tesla’s own operating instructions make clear that the human driver remains responsible for the vehicle.
Tesla’s Headline Safety Claim Is Seven Times Safer
Tesla’s most eye-catching Canadian safety message is that vehicles using FSD (Supervised) experience seven times fewer major and minor collisions than the estimated U.S. average. Its safety report relies heavily on vehicle telemetry and classifies a collision as FSD-related when the system was active at any point during the five seconds before the event. Tesla says that rule prevents a last-second disengagement from removing a crash from the FSD category and makes the comparison less vulnerable to selective counting.
The company separates “major” crashes involving airbag or other pyrotechnic restraint deployment from “minor” crashes that meet a defined change-in-velocity threshold without deployment. Tesla argues that its connected fleet gives it a consistent way to measure incidents across driving modes. At the same time, its methodology acknowledges assumptions when Tesla telemetry is compared with U.S. government crash databases built from different reporting systems. Tesla also does not publish injury outcomes in the same safety report, saying injury data are not consistently available through automated vehicle telemetry alone.
Researchers Say the Comparison Still Has a Baseline Problem
The central dispute is not whether Tesla has a large amount of driving data. It is whether the comparison answers the question consumers think it answers. Reuters reported in May that Tesla had compared airbag-deployment crashes in its own vehicles with a broader federal category that included tow-away crashes. A University of Michigan transportation researcher recalculated the comparison using a more closely matched airbag-to-airbag measure and found a substantially smaller advantage than Tesla’s earlier headline figure.
Researchers also pointed to differences between the vehicles and situations being compared. Teslas are newer than the average U.S. vehicle and commonly include modern crash-avoidance technology. Drivers can also choose when to activate FSD, potentially avoiding conditions they consider difficult. Those factors make it hard to isolate how much of any observed safety difference comes from FSD itself. Tesla has since revised parts of its methodology and now promotes a seven-times-safer figure, but outside researchers continue to question what that headline establishes on its own.
Europe Is Testing the Evidence, Not Just the Software
The newest challenge to Tesla’s evidence is unfolding in Europe, where the company is seeking broader regulatory acceptance. Reuters reported that seven traffic-safety researchers reviewed studies comparing FSD vehicles operated by Tesla’s professional engineering drivers with ordinary Teslas driven by customers. In the studies examined, the FSD sample was too small for a robust crash-rate comparison, so Tesla relied on surrogate behaviours such as hard braking, rapid acceleration, cornering, horn use and turn-signal patterns.
Some measures can indicate risk, but the researchers said others have weak or uncertain links to actual crash danger. Tesla’s European dashboard explains that the FSD data come from trained engineering operators instructed to minimize manual input, while the study language cautions that the comparisons should not be read as causal estimates. Reuters also reported that Tesla had not sought independent peer review of the methodology. Regulators are therefore assessing a sophisticated body of evidence that critics say still falls short of proving Tesla’s broadest safety claims.
European Regulators Are Not Reaching the Same Conclusion
Dutch regulator RDW has nevertheless moved farther than most European authorities. After more than a year of testing on tracks and public roads, RDW granted a national approval for FSD (Supervised) in the Netherlands and said that, when used correctly, the system can make a positive contribution to road safety. The regulator also stressed that it is not a self-driving system: the driver remains responsible and must stay in control. Tesla is seeking a path that could eventually allow wider use across the European Union.
The debate is still moving. Germany’s transport minister has publicly supported prompt approval, while other European governments have raised concerns about how the system behaves under local traffic rules. Reuters reported on October 7 that Tesla offered to use the name “Tesla Assisted Driving” in Europe as part of efforts to address concerns that “Full Self-Driving” overstates what the software can currently do. An EU-wide decision has been pushed back, making Europe a live test of both Tesla’s technology and the credibility of its safety case.
U.S. Regulators Are Investigating Specific Driving Failures
U.S. regulators are examining specific failure modes rather than arguing only over marketing language. In October 2025, the National Highway Traffic Safety Administration opened a preliminary evaluation covering an estimated 2.88 million FSD-equipped Teslas after reports of traffic-law violations. The opening record listed 58 incidents, including 14 crashes and 23 reported injuries, involving behaviours such as proceeding through red lights, entering opposing lanes and using turn-only lanes incorrectly. The agency said it would examine whether drivers received enough warning and opportunity to intervene safely.
A separate investigation became more serious in March 2026, when NHTSA upgraded its reduced-visibility probe to an Engineering Analysis covering an estimated 3.2 million vehicles. The agency said it was examining whether FSD properly detects degraded camera visibility and warns drivers with enough time to react in conditions such as glare and airborne obscurants. That record listed nine crashes, including one fatality. These investigations are not final determinations that every FSD-equipped Tesla is defective, but they show regulators testing how supervision works when the system behaves unexpectedly.
Canada Has Already Seen an FSD Safety Recall
Canada has already dealt with a related FSD safety problem. Transport Canada’s 2023 recall of certain Tesla vehicles equipped with FSD Beta said the system could perform manoeuvres that break traffic laws in some situations if the driver failed to intervene. Tesla addressed the recall with an over-the-air software update, while the Canadian notice emphasized that FSD Beta was an advanced driver-assistance feature and that the person behind the wheel had to remain attentive and ready to take control.
The recall involved an earlier software version, so it should not be treated as a finding about every capability in today’s FSD release. It does show that Canadian authorities have previously treated inappropriate automated manoeuvres as a safety issue. Transport Canada’s broader framework says vehicles commercially available in Canada currently span SAE Levels 0 through 2, where the human driver still monitors the driving environment. Provinces and territories separately govern licensing, insurance and rules of the road, meaning responsibility does not disappear when assistance software is engaged.
The Human-Factors Problem Does Not Vanish With Better Software
The hardest safety problem may be the relationship between capable software and a human who can become less engaged when the car appears to handle routine driving. Level 2 systems can control steering and speed together, yet they still require continuous monitoring and a driver ready to intervene. Research reviewed by the Insurance Institute for Highway Safety has found that regular users of partial automation can become more willing to perform non-driving tasks and may learn how attention-warning systems behave.
One 2025 naturalistic study followed 14 drivers using Tesla Autopilot for a month. Across more than 12,000 miles with the system engaged, researchers recorded 3,858 attention-related warnings and 16 lockouts, while signs of disengagement increased around alerts as participants became familiar with the safeguards. That work examined Autopilot rather than today’s FSD, so it cannot serve as a direct FSD safety score. It does illustrate why driver monitoring remains central whenever automation can perform many tasks but still depends on a human supervisor to catch sudden failures.
What the Debate Means for Canadian Subscribers
For Canadian owners, the practical takeaway is not that FSD has no safety value, nor that Tesla’s telemetry should be ignored. The system performs a broad range of driving tasks, Tesla collects extensive real-world vehicle data, and regulators continue to evaluate where supervised automation can reduce risk. The unresolved question is narrower but important: whether Tesla’s strongest public comparisons demonstrate the size of the safety advantage the company advertises, rather than a mixture of software performance, newer vehicles, driver selection, road type and other factors.
That makes the $99 subscription a convenience purchase wrapped in a safety debate. Drivers can decide whether the assistance is worth $1,188 a year, but the subscription should not be treated as permission to stop driving. Tesla’s own Canadian instructions say the opposite: remain attentive, avoid complacency and be prepared to take over. Until independent, exposure-matched research can separate the effect of FSD from the other variables surrounding its use, the most defensible conclusion is that the technology is capable, closely scrutinized and still fundamentally supervised.
