The next emergency intervention from an equipped Tesla may involve more than a warning or a sudden application of the brakes. Tesla has introduced Automatic Collision Evasion, a feature that can activate Full Self-Driving (Supervised) while a person is driving manually.
Introduced with September’s FSD v14.3.9 rollout, the feature gives the driving software a new safety role. But the details deserve care: Tesla’s original announcement described intervention when braking alone might not prevent a collision, while public release notes emphasize driver inattention and suspected accidental disengagement. That distinction matters, as do the hardware requirements and operating limits. This remains supervised driver assistance—not a guarantee that the car can escape every dangerous situation.
A Safety Feature That Can Switch On FSD Itself
Tesla’s AI team announced the new capability on September 4, 2026, describing situations in which FSD could activate without the driver deliberately starting it. The company identified an imminent collision that automatic emergency braking might not resolve, substantial driver distraction, and an apparently accidental disengagement of FSD. The important change is therefore not simply another improvement to how FSD drives when selected. It is permission for the system to engage during a manually driven trip, under specified circumstances.
There is an important distinction between FSD being enabled in the vehicle’s settings and actively controlling the car. Tesla’s documented collision-evasion setup requires FSD to be available and enabled, even though the driver is operating the vehicle manually. An owner who has temporarily taken over steering has not necessarily disabled the underlying feature. Understanding those separate states helps explain how a car can appear to be under manual control immediately before the software intervenes.
The Public Release Notes Are More Narrowly Worded
The rollout contains a detail that broad descriptions can miss. Tesla’s employee-build notes for software 2026.27.5 explicitly listed two trigger categories: an imminent frontal collision that braking might not prevent, and insufficient attention or an apparently unintended FSD disengagement. The published customer-build notes for 2026.27.6, dated September 8, instead describe the attention and disengagement cases without separately listing the frontal-collision scenario. The later 2026.27.11 notes, dated September 23, retain that narrower wording.
That difference does not, by itself, establish precisely which underlying functions Tesla enabled or disabled in every vehicle. Release notes are descriptions, not a complete technical specification. It does mean the original announcement should not be treated as proof that every eligible owner received every advertised intervention immediately. The sound conclusion is narrower: automatic FSD engagement reached customer software, but the exact capabilities must be checked against the installed version and the vehicle’s own instructions. A rollout headline cannot substitute for that information.
How the Announced Highway Evasion Would Work
An in-car Tesla manual description reproduced during the rollout sets out broader highway-evasion behaviour, going beyond applying the brakes in a straight line. In the documented frontal-collision scenario, the system assesses whether steering around another vehicle offers a better chance of avoiding impact. A manoeuvre could include moving into another lane or onto the shoulder, with braking or acceleration as required. This is the most consequential part of the original announcement: the software would be choosing a different path, not merely reducing speed along the existing one.
The documented conditions are restrictive. They include a controlled-access highway, a speed below 85 mph—approximately 137 km/h—a functioning steering system, and a surface not detected as wet or slippery. No pedestrians, cyclists or other vulnerable road users may be detected in the vehicle’s path. Those conditions describe a particular emergency scenario, not a general promise to dodge anything on any road. They also reinforce why the launch description and an individual vehicle’s delivered capabilities should not be treated as interchangeable.
Distraction Can Trigger a Different Kind of Intervention
The customer release notes describe a situation that is easier to picture than an emergency lane change: a driver reaches toward the back seat and stops paying sufficient attention to the road. Automatic Collision Evasion can engage FSD when it detects that lack of attention, or when it judges that FSD may have been disengaged unintentionally. The notes say the system can then continue driving. That is a different trigger from determining that a frontal impact is imminent.
The danger behind that example is well established. The U.S. National Highway Traffic Safety Administration says looking away for five seconds at 55 mph covers approximately the length of a football field. Tesla’s cabin camera already monitors driver attentiveness during supervised FSD operation. However, an intervention feature should not become permission to reach away, check a phone or stop watching traffic. Its potential value lies in providing a backup when attention fails—not making distraction an acceptable part of driving.
FSD Access and Vehicle Hardware Still Matter
Automatic Collision Evasion is not presented as a feature for every Tesla already on the road. The published rollout notes identify it with Hardware 4. They separately list an FSD v14.2 Lite update for Hardware 3 vehicles, without including collision evasion among that version’s features. Receiving an FSD update, therefore, does not automatically mean receiving this particular safety function. Owners of otherwise similar vehicles may have different capabilities beneath the same bodywork.
FSD also needs to be available and enabled; merely owning a Tesla is not enough. Tesla’s broader documentation cautions that feature availability depends on factors including market, vehicle configuration, purchased options, hardware and software version. The practical check is the release information and instructions displayed by the individual vehicle, rather than a social-media demonstration from another owner. A car’s installed software, supported hardware and enabled settings should all be confirmed before anyone assumes automatic engagement is available during a manually driven trip.
Why This Is More Than Automatic Braking
Automatic emergency braking tackles a familiar problem: a collision is developing ahead, and the driver has not slowed sufficiently. Tesla describes its braking system as a way to reduce impact speed and collision severity, while warning that it cannot prevent every crash. The broader evasion concept adds another decision—whether changing direction could help. That creates a more demanding task because an escape route must be assessed alongside the hazard directly in front of the car.
Existing braking technology provides an important benchmark. A 2017 study by Insurance Institute for Highway Safety researcher Jessica Cicchino found that forward-collision warning combined with automatic braking reduced rear-end striking crashes by 50 percent, and those involving injuries by 56 percent. Those figures concern established front-crash prevention systems, not Tesla’s new collision-evasion feature. They show why automatic braking remains valuable in its own right. Any claimed additional benefit from automatic steering or FSD engagement needs its own evidence, rather than borrowing the proven results of a different technology.
The Driver Still Has to Be Ready to Take Over
An important phrase in the customer release notes is that the feature can continue driving after it activates FSD. Owners should not picture only a momentary nudge of the steering wheel followed by an immediate return to the previous driving state. Recognizing that the car has changed modes becomes part of understanding the feature. The appropriate response is to restore full attention, assess the surrounding traffic and remain prepared to control the vehicle—not assume the emergency software has made further supervision unnecessary.
Tesla’s instructions remain explicit: supervised FSD requires an attentive driver who can take over at any time. The distinction also fits the National Highway Traffic Safety Administration’s explanation of Level 2 assistance, where a system may control steering and speed together but the person remains responsible for driving. An unexpected activation does not transfer that responsibility away from the driver. Learning the vehicle’s current alerts, limitations and takeover instructions while parked is considerably more sensible than trying to discover them during an emergency.
Evasive Steering Is Not Unique to Tesla
Steering assistance during an emergency is not a new idea across the automotive industry. Ford, for example, describes an Evasive Steering Assist feature that adds steering support after the driver begins turning to avoid a road user. Once the vehicle passes the obstacle, the system can also help as the driver steers back toward the lane. In that arrangement, the driver initiates the evasive move, and the assistance supports the steering effort rather than independently starting an ongoing driving mode.
Tesla’s announced approach differs in the way it connects emergency intervention with FSD activation. The distinction is not simply that a computer can influence the steering wheel; other systems already do that. It is that eligible Tesla software can engage a broader driving-assistance system during manual driving. Comparing these features requires attention to who starts the manoeuvre, what the system controls and what happens afterward. Calling every steering-related safety function equivalent would obscure the differences that matter most in a stressful moment.
Bad Weather Still Complicates the Picture
A clear view of the road remains essential. Tesla’s FSD documentation warns that heavy rain, snow, fog, poor visibility and other difficult conditions can affect performance. Its collision-avoidance guidance also identifies glare and obstructed cameras as potential problems. Those limitations matter because the software must recognize both a developing threat and the surrounding road environment. A safety feature cannot be assumed to compensate for camera views that have become blocked or unreliable.
The original highway-evasion description adds a specific restriction: that intervention is not intended for a surface detected as wet or slippery. That condition should not be confused with a blanket statement that every attention-related intervention has identical weather rules. For a driver facing a snowy commute or heavy spray behind a truck, the practical message is simpler. Keep cameras clear, respond to availability warnings and leave room to react. Emergency assistance is a backup, not a reason to carry dry-road expectations into poor conditions. A software update does not erase the importance of visibility and grip.
The Real Test Will Be Fewer Crashes
The strongest measure of this feature will not be how dramatic an intervention looks, but whether it reduces crashes and injuries without creating unacceptable new risks. Earlier research provides a reason to demand careful comparisons. Research released in 2024 by IIHS researcher Jessica Cicchino found no convincing evidence that the partial-automation systems studied prevented crashes beyond their underlying crash-avoidance features. That research examined BMW and Nissan vehicles, not Tesla’s new collision-evasion system, so it cannot establish whether this rollout will succeed or fail.
It does underline the difference between adding capability and demonstrating a safety benefit. Tesla’s release notes describe intended behaviour; they do not provide an independently measured crash-reduction rate for Automatic Collision Evasion. Meaningful evaluation would need to consider comparable driving conditions, missed hazards, unnecessary activations and driver responses. For now, the development is significant because it extends FSD’s role into manually driven trips. Its ultimate value will depend on safer outcomes—not the impression that the person behind the wheel is no longer needed.
