Will Any Country Ban Human Driving by 2028? We Fact-Checked It

Updated November 10, 2026 · 14-minute read · Seven documented cases · Every figure links to a source at the bottom

The short answer: no. As of today, no country has banned human driving, and in our searches we found no bill, ministry roadmap or credible report that schedules a ban for 2028. The date isn’t invented, though. 2028 appears in real plans to allow more automation. And the most heated political fight over robotaxis this autumn went in the opposite direction: a U.S. city tried to force a human back into the driver’s seat.

Key takeaways

• No ban exists or is scheduled. Not nationally, not in any large city we could find.
• The safety case is strong, but not unblemished. Waymo reports 82% fewer injury crashes than human drivers. An independent IIHS study found 68% fewer police-reportable crashes overall, with one city (Austin) slightly worse.
• The scale gap is enormous. Driverless miles are roughly 0.009% of one year of U.S. driving. Replacing the U.S. fleet would take about 18 years even if every new car were driverless.
• Politics runs both ways. In October 2026, Minneapolis passed, and then failed to rescue from a veto, an ordinance requiring human monitors in robotaxis.
• If a restriction ever arrives, it starts as a zone, such as a lane, a campus, or a downtown core, not as a national switch.

Claim check at a glance

In this article: where the “2028” idea comes from · how we checked · seven case files · the arithmetic of a 2028 ban · why voters matter · what a restriction would look like · what to do about it · signals to watch · FAQ · sources.

Where the “2028 ban” idea probably comes from

We couldn’t trace one origin, so treat this as informed reconstruction. Three true things get blended into one false one:

1. Germany’s federal strategy (adopted December 4, 2024) aims to create the world’s largest connected operating area for driverless vehicles by 2028. That is, to permit driverless service.
2. Carmakers’ 2028 targets. GM and Ford have both announced Level 3 “eyes-off” driving for 2028, and Stellantis has a hands-free, door-to-door supervised system aimed at North America for the same year. Those are optional features you buy.
3. A famous prediction. In 2015, Elon Musk told an Nvidia conference that people might one day outlaw driving because it is “too dangerous.” He added on Twitter that Tesla strongly supports people being allowed to drive and always will. He also said lawmakers would likely act two or three years after driverless cars became undoubtedly safer than humans.

That last point is our own hypothesis, so weigh it accordingly: large-scale, hard-to-ignore safety data only arrived in 2025–2026. Add Musk’s two-to-three-year lag, and 2028 becomes the year people might start debating a restriction. Debating is not enacting.

How we checked it (and what counts as a “ban”)

We searched for any national or state law, draft bill, ministry roadmap, or official statement proposing to prohibit human-operated cars by 2028 or any near date. We also looked at prediction markets, which are crude but show what informed hobbyists believe, then collected the strongest hard data on the other side: crash statistics, deployment numbers, regulatory timelines and public opinion.

For this article, a ban means a law that stops ordinary people from operating cars on public roads, either nationwide or in a defined zone. A permission lets driverless services operate. A mandate requires a human to be present. Headlines blur these three constantly, and most “ban” stories are really permissions or mandates.

Absence of evidence isn’t proof of absence. If you have a primary source for a real ban plan, we’d genuinely like to see it.

Case file 1: Waymo, the strongest argument for a ban

If any company can argue that humans should step aside, it is Waymo. Alphabet’s CEO said on the Q3 earnings call (October 28, 2026) that the fleet has passed 300 million fully driverless miles. The most recent detailed safety analysis, through June 2026, covers roughly 270 million.

Waymo’s own numbers

Each full bar = 100% reduction. Source: Waymo Safety Impact hub, data through June 2026, as reported by Electrek and EVwire.

The independent check

Company-reported numbers deserve a second opinion. In July 2026, the Insurance Institute for Highway Safety (IIHS) published its own analysis of crash reports from 2021 to 2024, comparing about 50 million Waymo driverless miles with roughly 222 billion human miles in the same places:

Sources: IIHS, Claims Journal, Carscoops.

Two details make this the most useful table in the article. First, the independent result is weaker than the company’s (68% is not 82%), which is exactly what you’d expect from stricter comparison methods, and it still points in the same direction. Second, IIHS flagged that about half of all human crashes, and a third of human injury crashes, never reach police records, so any comparison is imperfect. Its lead author also said the current federal data system isn’t good enough to monitor a large-scale expansion.

The part that complicates a ban

• School buses. NHTSA opened an investigation in October 2025 after footage of a Waymo passing a stopped school bus. Waymo issued a software recall in December 2025. Austin’s school district later counted at least two dozen incidents, some after the fix (TechCrunch, recall roundup).
• Floods. Waymo recalled its entire U.S. fleet, about 3,800 vehicles, after one drove into a flooded road in San Antonio and was swept into a creek. The recall was filed April 30, 2026 (TTNews, Automotive World).
• A child struck in Santa Monica in January 2026 (minor injuries) triggered another NHTSA preliminary evaluation. A third software recall, covering freeway construction zones, was reported in the summer.

Growth is real, but the headline target is unconfirmed

Sources: InsideEVs, FourWeekMBA summary of Alphabet’s Q3 call, Motley Fool.

The last official weekly number we could find is the 500,000-plus from April. The October earnings call didn’t update it, and Waymo’s own October post only says “hundreds of thousands.” Treat the one-million goal as an ambition until the company reports it.

What this case teaches: the safety argument is strong, independently supported, and still imperfect. Regulators are in tighten-and-fix mode, not replace-the-humans mode.

Case file 2: Cruise, or how one bad night ends a billion-dollar program

In October 2023, a Cruise robotaxi in San Francisco dragged a pedestrian about 20 feet after another car had thrown her into its path. She survived. What destroyed the company was what followed: California suspended its driverless permit, the state utility regulator alleged the company hadn’t been fully transparent for more than two weeks, nine executives were fired, a quarter of the workforce was cut, and in December 2024 GM stopped funding the unit. According to GM filings, Cruise piled up more than $10 billion in operating losses against under $500 million in revenue (KQED, AP via Click2Houston).

What this case teaches: the political risk runs against autonomy as easily as for it. One high-profile failure wiped out a leading competitor in months. A government that banned human drivers would be betting its whole road network on a technology whose public license can be pulled after a single incident.

Case file 3: Germany, first to legalize driverless cars, still not banning anything

Germany is the country people point to when they imagine a ban, so precision matters. It passed a law in 2021 allowing Level 4 vehicles, which need no human inside defined areas, to run in regular operation, backed by a 2022 operating ordinance and a requirement for remote technical supervisors. In December 2024, it adopted a national strategy: driverless vehicles integral to the mobility system by 2030, the world’s largest connected operating area by 2028, and an early focus on public transport and freight (strategy summary, law entering force).

Every verb there is about permission. The law builds lanes for machines and leaves steering wheels alone. It also describes itself as transitional until EU- and UN-level rules exist, reminding us how much international harmonization stands between any single country and a total ban.

What this case teaches: the most advanced legal framework on earth is a green light for driverless cars, not a red light for human ones.

Case file 4: The UK, where the permit nobody has finished

The UK’s Automated Vehicles Act 2024 places responsibility on the company, not the human, when the car drives itself. The government fast-tracked a permit scheme for driverless taxi-like services, and it has been legally open for trials since mid-May 2026. The fuller regime is expected in the second half of 2027 (HLC, Wikipedia summary).

The telling detail comes from The Next Web’s reporting: London’s first robotaxi service launched on ordinary private-hire licences with a human in the driver’s seat, and no operator had completed even the first stage of the driverless permit process in the capital (The Next Web). Waymo has said it plans a London service, but that route is gated on permits, vehicle approvals and local consent.

What this case teaches: even a country that wrote the rules in advance is moving city by city, permit by permit. That is the opposite of a ban switch.

Case file 5: China, scale without a ban

If any country could leap to a restriction, it would be one with a state-directed industrial policy and a huge driverless fleet. China’s Baidu reported that its Apollo Go service delivered 3.4 million fully driverless rides in Q4 2025 and 3.2 million in Q1 2026, with more than 22 million cumulative rides by April (Nasdaq summary of Baidu’s call). China’s industry ministry has said it supports commercial use of Level 3 and above systems (China Daily HK).

Then, on March 31, 2026, many Apollo Go vehicles stopped on the roads in Wuhan at once, which analysts described as a fleet-level failure rather than a single-car fault (The Next Web). Even with a state pushing the technology, it is regulating, not banning humans.

What this case teaches: scale amplifies both benefits and failure modes. A system that can strand a hundred cars at once is a system regulators will want to supervise, not take over everything.

Case file 6: Minneapolis, where the backlash ran the opposite way

This is the freshest and most relevant case. In late September and early October 2026, the Minneapolis City Council passed an ordinance requiring a paid, licensed human monitor in the driver’s seat of commercial robotaxis, with immediate access to steering, brakes, and accelerator. Supporters framed it as protecting about 10,000 rideshare drivers. Waymo called it a “de facto ban” on driverless service. Mayor Jacob Frey vetoed it, and on October 15 the council fell short of the nine votes needed to override, 7–5. Backers say they’ll try again, and neighboring St. Paul approved a study (Route Fifty, MPR News, KSTP).

What this case teaches: the first serious U.S. legislative fight over robotaxis asked for more humans, not fewer. Jobs, unions and distrust are a powerful counterweight to any “ban human driving” agenda, and the ordinance wouldn’t have taken effect until October 2027 anyway.

Case file 7: The 2015 prediction, eleven years later

In March 2015, Musk said driverless cars were nearly a solved problem and that people might eventually outlaw driving (Fortune). Eleven years on, driverless cars are real but geofenced, the U.S. fleet is a few thousand vehicles, and not one jurisdiction has outlawed human driving. Crowd forecasts mirror that long horizon: a Manifold market on whether any country bans human-driven cars in at least one city center by 2040 sits around 40% (thin and play-money, so treat it gently), and the Metaculus community median for two U.S. states barring human-driven vehicles is around 2068, with a range from the early 2050s to the 2090s.

What this case teaches: bold timelines from the people building the technology run early by years, and forecasters who do this for a living put real prohibitions decades out.

The arithmetic of a 2028 ban

We built this table because we couldn’t find it anywhere else. The U.S. is the test case: biggest fleet, best data. Rows marked “our math” are simple division from the cited figures.

So even in the fantasy where every new car sold from tomorrow had no steering wheel, human-driven cars would still be a majority of vehicles until the mid-2040s. A 2028 ban would mean making roughly 286 million vehicles illegal to operate, or fencing drivers into a few designated roads, in about two years.

The part everyone skips: voters have to agree

A ban is a law, and laws need votes. The latest national AAA survey we could verify (February 2025) found only 13% of U.S. drivers would trust riding in a self-driving vehicle, with 61% afraid and 26% unsure. 78% said improving safety systems matters more than developing self-driving cars, and 53% wouldn’t choose to ride a robotaxi (AAA Oregon/Idaho). Minneapolis shows how those attitudes turn into ordinances.

The stakes people feel now are real, too. U.S. traffic deaths were about 39,254 in 2024 and an estimated 36,640 in 2025, a 6.7% drop, with the death rate falling to 1.10 per 100 million miles (NHTSA, AASHTO Journal). Globally, the WHO counts 1.19 million road deaths a year. Those figures are the moral engine behind any future restriction, and because they’re falling without a ban, the urgency argument is weaker in the short term.

One more reality check: the vehicle “ban” bill that actually advanced in Congress this year, the Connected Vehicle Security Act, targets internet-connected vehicles tied to foreign adversaries, not human drivers (overview).

If a restriction ever comes, here’s the likely shape (our analysis, not a forecast)

Permissions flip before prohibitions. Based on how cities handled low-emission zones and bus lanes, the realistic ladder looks like this:

1. Permit driverless taxis city by city (we are here).
2. Prioritize them: dedicated lanes, airport and campus zones, preferred curb access.
3. Zone them: districts or highways where only automated vehicles may operate at peak hours.
4. Restrict manual driving in those zones, with exemptions for emergency vehicles, disabled drivers, and permit holders.
5. Prohibit broadly, if ever, after the older fleet has aged out.

A first headline-making “ban” will almost certainly be step 3 or 4 in a small area: a campus, a downtown core, a freight corridor. If a story calls that “a ban on human driving,” you’re looking at a zone.

What to do with this

If you drive for a living (taxi, rideshare, trucking). Your risk is competitive, not legal. Track the permit list for your metro rather than the legislature, and watch whether your city is debating a Minneapolis-style human-monitor rule, because that is where jobs are being decided. Driverless services create jobs in fleet supervision, remote assistance, and logistics.

If you’re a regular driver. Nothing about your right to drive changes in 2028. Look up your own car on the NHTSA recall lookup, because defects and distraction are the live safety issues today. If you ride a robotaxi, remember the data says it’s safer in its service area, not perfect, and be extra careful at school zones.

If you invest or advise. Don’t price a ban into anything. Price permit-by-permit expansion, recall and liability risk, and the fleet-math limits on growth. We’re not financial advisors; this is a framing, not a recommendation.

If you write or share news. Use a four-step test on viral claims like this one: (1) find the primary document, such as a bill, a ministry page or a court order; (2) check whether the verb is allow, require or prohibit; (3) check that the date is a legal effective date, not a company goal; (4) find a second independent source. The “2028 ban” claim fails step one.

If you’re a policymaker or local official. Copy the useful parts of the UK and German approach (permits, safety cases, local consent) and the useful part of the IIHS warning: demand mileage and crash data from every operator, because only one company reports it voluntarily today.

Five signals that would change our minds

We’ll update this article if any of these appear:

• A government publishes draft legislation using the word prohibit for manual operation, even in a limited zone.
• A transport ministry publishes a dated roadmap for phasing out human-operated vehicles.
• A national insurance regime starts pricing human driving as the exceptional, higher-risk case.
• A large jurisdiction passes a law reserving a named road or district for automated vehicles only.
• A major operator completes a full year of expansion with no safety recall or regulatory action.

FAQ

Is any country planning to ban human driving?

Not that we could find. Germany, the UK and China are all writing rules to allow driverless vehicles. None has proposed prohibiting human-operated cars.

Will human driving ever be banned?

Possibly in limited zones, and much later than 2028. Forecasting communities put broad prohibitions decades away, and the vehicle fleet turns over too slowly for a near-term ban.

When could driving yourself become illegal?

No official date exists. The most plausible first step is a restricted lane or district for automated vehicles, not a nationwide rule.

Are self-driving cars really safer than human drivers?

In the cities where they operate, the evidence says yes. Waymo reports 82% fewer injury crashes, and an independent IIHS study found 68% fewer police-reportable crashes. The data is limited to one main operator and geofenced areas, and recalls show the technology still has failure modes.

Do I need to give up my car?

No. Nothing in current law or any proposal we found requires it.

What is a Level 4 vehicle?

A car that drives itself without human supervision within a defined area and set of conditions, such as a robotaxi in a mapped city. Level 3 still needs a human ready to take over when asked.

The bottom line

The evidence backs a surprising conclusion: the safety case for driverless cars keeps getting stronger, and the case for a human-driving ban by 2028 keeps getting weaker. The first is about miles of data. The second is about fleets, politics, and trust. The most plausible near-term future isn’t a country forbidding you to drive. It’s a quiet, city-by-city handover, with more driverless taxis, more restricted zones, and loud local fights over whether a human should be required. The steering wheel stays legal almost everywhere.

Laws that take freedoms away move slower than technology that offers convenience. If a ban comes, it will arrive long after the headline has been forgotten.

Sources

Safety and deployment

• Waymo Safety Impact hub and September 2026 safety update
• IIHS: Waymo’s driverless cars crash less often than people
• Claims Journal on the IIHS study · Carscoops
• Electrek on Waymo’s 841 avoided injuries · EVwire
• InsideEVs: Waymo weekly rides · FourWeekMBA on Alphabet’s Q3 call · Motley Fool
• TTNews: flood recall · Automotive World · TechCrunch: school bus recall · Recall roundup

Cruise

• KQED: GM ends Cruise robotaxis · AP via Click2Houston

Regulation and politics

• Germany’s 2024 strategy · German Level 4 law
• UK Automated Vehicles Act in 2026 (HLC) · Act overview · The Next Web on London’s permit gap
• Baidu Q1 2026 call (Nasdaq) · China Daily HK on Level 3 policy · The Next Web on the Wuhan stall
• Route Fifty: Minneapolis veto · MPR News · KSTP · CBS Minnesota
• Automaker 2028 targets (CBT News) · Connected Vehicle Security Act

Road deaths and the fleet

• WHO Global Status Report on Road Safety 2023
• NHTSA early estimate for 2025 · AASHTO Journal
• S&P Global Mobility via Automotive Fleet · via Auto Dealer Today

Opinion and forecasts

• AAA autonomous vehicle survey (Feb 2025)
• Fortune on Musk’s 2015 comments
• Manifold market · Metaculus question

Methodology and corrections

Waymo’s figures are company-reported; the IIHS figures are independent but cover a different period and metric, so the two sets shouldn’t be mixed. Rows marked “our math” are our own division of the cited numbers and are approximations. The U.S. fleet figure dates from January 2024 and has likely grown, which makes our scale comparisons conservative. This article is informational and is not legal, financial, or safety advice. Spotted an error or have a primary source we missed? Tell us, and we’ll correct it and date the change.