TheChek News Flow

NHTSA investigates Ford’s product after deadly crashes; company research during development showed technology can be misunderstood

A Ford engineer and her husband were driving their new F-150 pickup in May when something went terribly wrong. The engineer told Ford they were using Ford’s hands-free driving system, BlueCruise, which puts computers in charge of the truck’s speed and steering.

Just outside downtown Toledo, Ohio, on I-75, Seetaram Palepu tapped the brakes to slow as he merged onto an exit ramp. But instead, the truck accelerated. It smashed into a guardrail, rolled over and came to a stop upside down. The Palepus had to be extracted by emergency workers, according to the bodycam footage of responding officers.

The truck, Palepu later said in a court filing, “went out of control.”

Ford is a leader in the industrywide push to automate driving. Its BlueCruise driving system promises to keep a car in its lane, manage its speed and slow it down upon approaching another vehicle, without the driver’s hands on the steering wheel.

More than two million Ford cars and trucks in 2025 were equipped with the technology or some of its elements, and Chief Executive Jim Farley said the automaker sees a lot of potential in developing an even more advanced automated driving system that allows for “eyes off” driving.

But some drivers have found the technology behind BlueCruise confusing to use, according to driver statements, crash reports and Ford’s research. Some haven’t understood the product’s limitations, while others haven’t heeded the system’s warnings to take control of the car—or haven’t had time to.

Ford has long known about the dangers. Confidential company documents from 2018-19, reviewed by The Wall Street Journal, show Ford learned while developing BlueCruise that some drivers didn’t understand how to use certain features, stopped paying attention to the road or failed to respond to warnings to retake control.

Ford told the Journal it took steps to improve BlueCruise before rolling it out, such as by refining graphics on the vehicle dashboard and the messaging used to alert drivers, “to further improve the driver’s understanding of their responsibility.”

The full BlueCruise product, which allows hands-free driving, was released in 2021, although some features of the system, such as lane centering and adaptive cruise control, were available earlier.

But after two deadly crashes in 2024, federal regulators are looking into whether the technology is still posing problems for drivers.

Since 2021, Ford has reported to U.S. regulators 32 crashes, including three that were fatal, in which the company’s automated driving technology was suspected to be engaged at the time of the collision, according to federal data.

Ford told the Journal it has built extensive safeguards into BlueCruise to ensure drivers use it correctly and are ready to resume control of the vehicle, including an eye-tracking camera that will alert drivers if they look away from the road.

Customer feedback and company research haven’t found any signs that drivers find the technology confusing since it launched, the automaker said. If drivers pay attention, Ford’s automated-driving technology is safer than driving unassisted, the company said.

“We have safely enabled more than half a billion miles of travel by helping drivers remain engaged,” Ford said.

Automating driving has proven challenging for many companies. The National Highway Traffic Safety Administration, the top auto-safety regulator in the U.S., is probing Tesla’s “Full Self-Driving” system and the driverless robotaxis operated by Alphabet-owned Waymo. Waymo said it would cooperate with NHTSA.

After two cars using Ford’s BlueCruise hit stopped cars in 2024, NHTSA began investigating fatal crashes involving the system. The agency said it has received more than 2,000 reports from drivers saying they experienced an issue with BlueCruise or Ford’s other automated-driving features.

NHTSA said last year that “system limitations” detecting stationary vehicles while driving at night and at highway speeds appeared to be factors in the collisions it was investigating and several near misses.

Ford said to the Journal that roughly three dozen of the 2,000 reports to NHTSA involved crashes, of which nine involved the hands-free BlueCruise feature.

Padmalaya Palepu, the Ford engineer who was in the truck’s passenger seat while her husband drove, told Ford her husband had tried disabling BlueCruise before the crash because their 2025 F-150 pickup started drifting out of its lane, but the system wouldn’t turn off, according to the initial report Ford submitted to NHTSA relaying Padmalaya’s claims.

Later, the company told NHTSA that data from the vehicle showed Seetaram had turned off the F-150’s cruise control before colliding with the guardrail and had fully compressed the gas pedal.

Neither BlueCruise nor the cruise-control function were active during the 20 seconds prior to the crash, Ford said to the Journal. “Any confusion involved in the accident was related to the driver’s pedal application and not the BlueCruise or” cruise-control system, Ford said.   

Padmalaya Palepu, whose LinkedIn says she is an analytical design manager at the automaker, declined to comment through Ford, and Seetaram Palepu didn’t respond to requests for comment.

Both Seetaram and Padmalaya were treated at a hospital after the crash. Seetaram required surgery and wore a neck brace, according to court records and police filings.

Police issued Seetaram a traffic ticket for failing to control the vehicle. In late May, he told a Toledo court he wanted to contest the citation because of the F-150’s driving technology. “Accident happened as the vehicle went out of control, due to cruise-control malfunction,” he said in a court filing. He later paid a fine for a reduced charge, records show.

NHTSA declined to comment on the Palepu incident because of the ongoing investigation. Jonathan Morrison, the agency’s administrator, said at a January conference that drivers should know they need to remain in the loop when using these systems. “We want to be crystal clear that the systems available in today’s vehicles are driver assistance systems: you’re driving,” he said.

‘Common areas of confusion’

Since Tesla released its Autopilot system in 2015, carmakers have been locked in a fierce arms race to advance automated-driving technology. Many have sought to proceed in steps; they’ve rolled out automating braking, advanced cruise control and, for some, hands-free driving. Some aim to develop cars that would drive themselves—technology that currently is mostly limited to robotaxis operating in several cities.

Ford spent years saying partially automated systems were too much of a challenge, because drivers must remain attentive and ready to assume control of the wheel, and executives said some drivers can become complacent while using the technology.

But amid the technical challenges and costs behind making full self-driving technology, Ford refocused in the late-2010s on developing a driver-assistance system controlling steering and speed—the product that eventually became BlueCruise.

In June 2018, Ford launched a benchmarking study using driver-assistance technology from General Motors. Ford looked at its rival’s technology to determine a design direction for its own system, according to a company presentation.

Ford enlisted more than 40 people to drive a 2018 Cadillac CT6 with GM’s Super Cruise along a highway near Grand Rapids, Mich., and asked them about the experience. Their responses confirmed the company’s earlier misgivings: “Drivers worry they will be too complacent or lazy behind the wheel with the system active (worry about themselves and others),” the presentation said.

The presentation is among test reports, prototype studies and other documents that Ford submitted to NHTSA in response to the agency’s BlueCruise investigation. The documents are available to the public with redactions. The Journal reviewed hundreds of pages of the documents without the redactions.

After the test drives were completed, Ford identified a lengthy list of “common areas of confusion and concern” while using GM’s automated-driving system.

If the technology’s auto-steering fails and veers the car out of a lane, drivers are supposed to grab the wheel and reposition it. Yet one-in-four drivers incorrectly believed Super Cruise would handle the function, the presentation said.

Also, 80% failed to notice an initial flashing green light warning drivers to resume control of the wheel. Drivers later noticed additional warnings.

A GM spokeswoman told the Journal that references to the 2018 study don’t “reflect our current system or the familiarity customers now have with hands-free technology,” adding that the GM system “has been validated across 800 million real-world miles.”

In a 2019 study of its own developing product, in which Ford sent drivers along a stretch of Highway 101 in California, the company found more than half the drivers failed to engage the correct adaptive cruise-control button, and 60% of drivers incorrectly believed the technology would reposition the vehicle if it veered out of a lane. “It’s still not clear to the driver that they are responsible for steering the [vehicle] back into the lane,” a presentation about the study, reviewed by the Journal, said.

When Ford simulated an event that required drivers to take control of the wheel, 10% of them didn’t respond to the initial warning prompts from the system, according to the presentation.

The study, however, showed that novice users quickly improved their understanding of the system’s functionality, Ford said. Nearly 100% of subjects in pre-launch testing demonstrated an “understanding of critical components of the driver assistance features after a single drive session,” the company said.

Based on the findings, Ford said it made improvements to the system before launching the first product named BlueCruise in 2021.

‘Black thing’ in the road

In September 2021, Barry Wooten was killed in a collision in Forsyth, Ga., after losing control of his F-150 truck in “self-driving mode,” according to a report the driver’s family filed with NHTSA.

Wooten’s family said in court filings he was using elements of the BlueCruise system at the time of the incident. The family has filed a personal-injury lawsuit against Ford.

Wooten spent his entire life in the car business, said his daughter, Wendy Wells. He was a national director of service and parts at a Georgia auto dealer chain that sold Ford cars, among other brands. “Dad wanted the first of everything,” she said. “He wanted the first new body style Corvette. He loved technology and having the latest and greatest.”

While driving on a highway in Georgia, Barry Wooten’s F-150 plotted a path for it to drive off-road as he attempted to take an exit ramp, the family said in court documents.

Based on GPS location and data taken from the car at the time of impact, Wooten didn’t have enough time to realize the truck wasn’t properly making the exit and to take control of the vehicle, the family’s filings said.

Ford said it doesn’t comment on pending litigation. Lawyers for the Wooten family declined to comment because of the pending case.

In a report to NHTSA that covered several incidents, Ford said Wooten’s F-150 was “not equipped with the operating system necessary to make BlueCruise functional,” meaning it didn’t have the feature that allowed hands-free driving. In other documents, Ford refers to the kind of technology equipped on Wooten’s F-150 as “hands-on” BlueCruise.

“Ford’s assessment of this incident is that the subject system performed as designed and that there was no failure of the subject system in the subject crash,” the company said in another report.

In 2024, on a February night, an all-electric Ford Mustang Mach-E SUV driving on a highway in San Antonio struck the rear of a Honda CR-V that was stopped in a traffic lane, according to police and federal crash reports.

The Ford driver told police he didn’t have time to react after he “saw a black thing” in the road moments before the crash, the federal crash report said. The driver also said the Honda was at a “complete stop with no lights on” at the time, the police report said.

The Honda overturned, and the impact was so severe the Honda’s driver was fatally injured.

Ford said in a report to NHTSA that BlueCruise was engaged at the time but that warnings to take control of the wheel were ignored for 30 seconds prior to the collision.

One night a couple of weeks later, in March 2024, another Mustang Mach-E owner using BlueCruise in a construction zone on a Pennsylvania highway slammed into a stationary Hyundai Elantra sedan. The impact pushed the Elantra into a Toyota Prius, and both the Prius and Elantra drivers died at the scene.

The Prius and Elantra appeared to have been involved in an earlier collision, a crash report said. The cars were in the highway’s left lane, and the Prius driver was standing next to the Elantra. As the Elantra driver began exiting the vehicle, the Mustang Mach-E slammed into the car, according to the Mustang’s camera and other information.

Police later charged the Ford driver with vehicular homicide and found she was drunk at the time of the collision.

Based on images from the Mustang’s system, Ford said in a report to NHTSA the driver may have been attempting to defeat the eye-tracking camera with an electronic device and had ignored warnings for nearly 30 seconds before the crash.

The Ford driver pleaded not guilty, according to her attorney, Zak Goldstein. A trial date hasn’t been set.

NHTSA formally decided to examine whether BlueCruise had a defect after the pair of deadly crashes.

Ford told the Journal that neither accident was caused or contributed to by Ford’s automated cruise-control design. The two accidents, the company said, unfortunately illustrate that all driving, whether by humans alone or with technology, requires adequate time to perceive, classify, confirm and react to events.

After examining data from the vehicles involved in the 2024 crashes, NHTSA said the drivers in each case failed to apply brakes or take evasive steering action. NHTSA also said BlueCruise’s adaptive cruise-control function failed to detect or respond to stopped or slow-moving vehicles.

Ford told the agency that it designed the system’s adaptive cruise control to stop decelerating in response to stationary objects when traveling at or above 62 miles an hour—to avoid the “phantom braking” phenomenon, in which the vehicle incorrectly reacts to objects such as bridges, overhead signs or other nonmoving objects that aren’t in the roadway. Ford said that other technology, such as automatic emergency braking, remains functional at speeds above 62 mph.

“Ford has designed the feature to work so that it can manage various real-world scenarios that can occur on various road types,” a spokeswoman told the Journal. The company also said there may not be adequate time for either a human driver or the automated driving system to respond under certain conditions, such as approaching an unlit vehicle at nighttime at highway speeds.

Farley, the CEO, told analysts early last year that Ford sees a lot of potential in developing a robust automated driving system that allows for “eyes off” driving in some scenarios.

Farley said current products from other carmakers only work at low speeds, and that Ford wants to offer a “fully functioning” high-speed system so that millions of customers could use it on the highway and do other tasks. The company said they want to introduce such a system as soon as 2028.

The vote is a stinging rebuke of GOP leadership that paves the way for challenges to Trump’s signature economic policy

WASHINGTON—House lawmakers on Tuesday rejected an attempt by Speaker Mike Johnson (R., La.) to block votes on resolutions disapproving of President Trump’s tariffs—a stinging blow to his leadership that paves the way for lawmakers to potentially rebuke Trump’s signature economic policy.

The procedural step failed with 217 opposed and 214 in favor, with three Republicans joining all 214 Democrats in voting against the measure, enough to sink it in the narrowly divided chamber.

The no votes came from Republicans across the ideological GOP spectrum: centrist Reps. Don Bacon of Nebraska and Kevin Kiley of California, as well as libertarian Rep. Thomas Massie (R., Ky.). The vote was kept open for about an hour as leadership looked to flip votes, but none of the defectors budged. Rep. Greg Murphy (R., N.C.) didn’t vote.

The vote means that Democrats will be able to bring resolutions challenging Trump’s tariffs to the House floor, setting up a series of high-profile votes that could begin as soon as Wednesday. Though Trump could veto any measure that reaches his desk, any successful vote would be a public repudiation of his tariff policy and would likely draw a furious reaction from the White House.

“Big step forward for Americans tired of paying more because of Trump’s tariffs,” said Rep. Suzan DelBene (D., Wash.) after the vote.

Johnson was trying to extend a moratorium on any tariff votes in the GOP-controlled chamber. Last year, he employed a procedural maneuver to block Democrats from bringing up votes on resolutions that would invalidate the emergency declarations that underpin many of the levies. Similar resolutions against his tariffs on Canada, Brazil and other global trading partners passed the Senate with Democratic and some Republican votes last year.

The timeline on Johnson’s blocking maneuver expired this month, prompting Democrats to ready a resolution disapproving of Trump’s tariffs on Canada, which the administration says are based on a national emergency stemming from the fentanyl trade, despite little of the drug actually coming from Canada.

Johnson on Tuesday tried to block the tariff resolutions again, putting a rule on the House floor that would have prevented any votes on them until August. But he was foiled by his razor-thin majority—now just 218-214 when all lawmakers are present—and the handful of pro-free-trade Republicans who have criticized Trump’s aggressive use of tariffs and their questionable legal underpinning.

Johnson, in a brief interview, said GOP leaders postponed the scheduled vote from Tuesday afternoon to Tuesday night to give him time to work to flip Republican votes. U.S. Trade Representative Jamieson Greer also met with some House members in the Capitol.

Republican leaders wanted to push the extension past the deadline of when the Supreme Court is expected to rule on the president’s authority to levy emergency tariffs. The ruling is expected by the end of June, when the court’s session ends.

“I think the sentiment is that we allow a little bit more runway for this to be worked out between the executive branch and the judicial branch, and we’ll see,” Johnson said at a press conference Tuesday morning.

Bacon listed two main reasons for his opposition: that Congress should retain its authority to levy taxes including tariffs and that these particular taxes are “generally bad policy.”

“I have a heavy heart opposing the majority here, but Republicans have opposed tariffs since WW2. I have not changed. I follow a moral compass, not a person, when it comes to key policies,” said Bacon, who is retiring at the end of this term.

Democrats could bring up resolutions challenging Trump’s tariffs as soon as Wednesday. The first will be a vote to disapprove of Trump’s tariffs on Canada, Democratic aides said this week, and could be followed by resolutions targeting tariffs against Brazil and other nations.

Though the Senate has already passed similar resolutions last year, any House-passed measures would still need to be considered by the Senate before going to the president. Congress is unlikely to have the votes to override a presidential veto.

The executive, who was accused of sexual discrimination against a male employee, had raised concerns about upcoming launch of erotic content

OpenAI has cut ties with one of its top safety executives, on the grounds of sexual discrimination, after she voiced opposition to the controversial rollout of AI erotica in its ChatGPT product.

The fast-growing artificial intelligence company fired the executive, Ryan Beiermeister, in early January, following a leave of absence, according to people familiar with the matter. OpenAI told her the termination was related to her sexual discrimination against a male colleague.

“The allegation that I discriminated against anyone is absolutely false,” Beiermeister said in a statement in response to a request for comment.

OpenAI said Beiermeister “made valuable contributions during her time at OpenAI, and her departure was not related to any issue she raised while working at the company.”

Beiermeister served as the vice president leading OpenAI’s product policy team, which develops rules for how people can use the company’s products and helps design the enforcement mechanisms for those policies.

Her ousting came ahead of OpenAI’s planned launch early this year of a mode that will allow users to create AI erotica in ChatGPT. The planned feature, which would permit adult-themed conversation including sexual topics for adult users, has drawn criticism from researchers at the company who have studied the ways some people develop unhealthy attachments to chatbots, according to some of the people. They have raised the prospect that sexual content could intensify the feelings some people have for the AI personas they view as companions.

Members of an advisory council on “well-being and AI” that OpenAI convenes regularly have also expressed opposition to adult mode and urged the company to reconsider plans to launch it, people with knowledge of those discussions said.

OpenAI Chief Executive Sam Altman has defended the move to expand the permissible content on his platform as a part of an effort to “treat adult users like adults.”

Before her firing, Beiermeister told colleagues that she opposed adult mode, and worried it would have harmful effects for users, people familiar with her remarks said.

She also told colleagues that she believed OpenAI’s mechanisms to stop child-exploitation content weren’t effective enough, and that the company couldn’t sufficiently wall off adult content from teens, the people said.

She is one of several employees within OpenAI who have expressed concerns about the launch of adult mode, people familiar with the matter said.

OpenAI has drawn more than 800 million users each week to ChatGPT in its campaign to build the most advanced artificial intelligence. The company now plans to monetize the engagement of those users with advertising. News Corp, owner of The Wall Street Journal, has a content-licensing partnership with OpenAI.

OpenAI has kept close tabs on competitors, and in December declared a “code red” after the surprising success and growth of Google’s Gemini chatbot.

Another competitor, xAI, has found that offering looser guardrails around sexual content in its Grok chatbot has helped drive engagement.

Beiermeister started at OpenAI in mid-2024 as a part of a wave of hires from Meta who viewed themselves as trying to change tech companies from the inside, people with knowledge of the matter said.

Beiermeister started a peer-mentorship program for women at OpenAI in early 2025. The program connected women from different parts of the company and helped them gather in small groups to discuss career strategies, according to people familiar with the matter.

Activist Investor Pushes Warner to Walk Away From Netflix Deal

Ancora builds a roughly $200 million stake in Warner Bros. Discovery and favors a deal with David Ellison’s Paramount Skydance

Activist investor Ancora Holdings has built a roughly $200 million stake in Warner Bros. Discovery WBD 2.17%increase; green up pointing triangle and is planning to oppose Warner’s deal to sell its movie and television studios and HBO Max streaming service to Netflix NFLX 0.91%increase; green up pointing triangle, according to people familiar with the matter.

Ancora, which could announce its position as soon as Wednesday, believes that Warner failed to adequately engage with David Ellison’s Paramount Skydance PSKY 1.50%increase; green up pointing triangle after it made a rival all-cash offer for the entire business, including its cable-network group, at $30 a share, the people said.

The arrival of an activist, even with a small stake in the company, will add yet another dose of uncertainty and drama to an already drawn-out fight for the Hollywood studio. Netflix has signed a $72 billion deal, but Paramount, which is bidding nearly $78 billion for the whole company, has gone straight to shareholders and threatened to wage a board fight at the same time.

Ancora, a roughly $11 billion fund that has a history of lobbying in the middle of deals, emailed Warner Chief Executive Officer David Zaslav on Tuesday to say that it is considering launching its own proxy fight if Warner’s board doesn’t negotiate the best deal for shareholders with Paramount, the people added.

Warner has a market value of roughly $69 billion as of Tuesday, making Ancora’s stake in the company less than 1%. But Ancora plans to continue buying Warner shares, the people familiar with the matter added, and, even with a small stake, it adds a voice that could help rally other investors around opposing the Netflix transaction.

Many shareholders remain on the fence over which deal is better and are anticipating the offers could be revised further. A shareholder vote is expected next month.

Netflix agreed in December to pay $27.75 a share in cash for Warner’s studios and HBO Max streaming service. That would leave investors also holding shares in Discovery Global, a new company housing Warner’s cable networks, which it plans to spin off later this year. Paramount’s hostile bid for all of Warner Discovery includes its cable-networks unit that includes CNN, TNT, Food Network and other channels.

Warner has consistently rebuffed Paramount’s offer, arguing Netflix’s deal has greater value, more secure financing and a cleaner path to regulatory approval.

Paramount on Tuesday enhanced its hostile offer, including agreeing to pay the $2.8 billion termination fee Warner would owe Netflix should that deal collapse. Paramount also said it was adding a “ticking fee” of 25 cents a share, which it would pay to Warner shareholders for each quarter its deal hasn’t closed, starting in January 2027.

If Ancora were to proceed with nominating director candidates, it would focus on replacing individuals with ties to Zaslav, the people familiar with the matter said. Ancora has privately questioned the Warner CEO about whether he favored the Netflix deal to obtain an executive role with the streaming company after the transaction closes, they added.

Ancora has antitrust concerns about the Netflix deal it calls “uncertain and inferior.” And it questions the Discovery Global spinoff, which would put $17 billion in Warner debt on the company’s cable-TV networks, which have a declining number of viewers, according to a presentation from the activist seen by The Wall Street Journal.

In that presentation, Ancora defends Paramount’s viability as a buyer, pointing to the record of Ellison and his father, the billionaire Oracle co-founder Larry Ellison. It also said it expects Paramount to receive swift antitrust approval.

Many investors and analysts still largely expect Paramount could increase its $30-a-share offer. Analysts at Raymond James said in a recent note to clients that “many WBD shareholders still expect PSKY has not made its best and final offer, and will raise its bid by ~$2-3 per share.”

Cleveland-based Ancora has a history of pushing for deals, both publicly and behind the scenes. In 2024, it built a huge stake in Norfolk Southern and later won seats on the railroad operator’s board before the company agreed to be acquired by Union Pacific for almost $72 billion. It also recently privately pushed bubble-wrap maker Sealed Air to sell itself, before the business agreed to be bought by CD&R.

Ancora Holdings accused WBD board of failing to give proper consideration to rival bid from Paramount

Activist investor Ancora Holdings has taken a $200mn position in Warner Bros Discovery and intends to oppose the Hollywood studio’s agreed $83bn sale to Netflix.

Ancora is expected to argue that WBD’s board has not given adequate consideration to a rival bid from Paramount, according to people familiar with the matter. The company could outline its case in a filing as soon as Wednesday.   

The group’s stake in WBD is less than 1 per cent but Ancora is expected to buy more of the shares. It is also considering launching a proxy battle to enlist support from other investors. 

It marks the latest twist in the battle to land one of Hollywood’s biggest deals, as David Ellison’s Paramount has continued to try to win WBD away from Netflix.

Ancora has previously waged activist campaigns against industrial groups. It opposed US Steel’s agreement last year to sell to Nippon Steel, which ultimately was approved. It also pushed for transport group CSX to pursue a merger following a deal between railways Union Pacific and Norfolk Southern.

Paramount on Tuesday sweetened its $108bn hostile bid for WBD by offering shareholders an additional fee to compensate them if regulators delay completion of the transaction.  

Under the new proposal, WBD investors would receive an extra $0.25 a share — about $650mn — in quarterly payments should the deal fail to close by the end of 2026. Paramount said its enhanced proposal highlighted its confidence in securing approvals.

Paramount stopped short of raising its offer for WBD — a move that analysts say is necessary to sway more shareholders. WBD’s board last month rejected Paramount’s bid, calling it “inadequate”.  

WBD plans a shareholder vote on the Netflix deal by early spring.

If Paramount decides to increase its bid, it would probably do so before the shareholder vote, said people briefed about the matter.

Netflix is facing a US government antitrust review into the WBD deal. Last week co-chief executive Ted Sarandos faced questions in the US Senate over the deal.   

The Wall Street Journal first reported Ancora’s plans. WBD declined to comment. Ancora did not respond to a request for comment.

The shingles vaccine may have a dementia upside

Research suggests it could prevent and slow the progress of cognitive decline

The shingles vaccine could, at a stretch, be labelled an accidental blockbuster. It does its intended job of fending off the varicella-zoster virus, which causes both chickenpox and shingles, but a growing body of evidence hints it also protects against dementia, particularly among women.

Over the past year, data from separate vaccine rollouts in Wales, Australia, Canada and the US suggests it can delay the onset of dementia, slow down its progress and cut the risk of death among those already diagnosed. Despite the vaccine being aimed at the over-50s, the 39-year-old researcher leading some of the analyses is so convinced by the data that he has had the jab himself.

Pascal Geldsetzer, assistant professor of medicine at Stanford University, is now trying to set up a clinical trial to prove it is cause and effect, not mere correlation. “It’s extremely exciting because this is an inexpensive one-off intervention, not a medication that has to be taken every day or an exercise and diet regimen you have to stick to for decades,” he told me.

The unexplained connection between shingles vaccination and lowered dementia risk matters for several reasons: it highlights the power of “natural studies”, which harness mass observational data; it shows how repurposing vaccines and therapies can offer hope where novel offerings are falling short; it implicates viruses in neurodegenerative diseases; and it reveals the potential wider gains of vaccination at a time when public confidence is wobbling.

The studies focus mostly on the Zostavax vaccine, a live attenuated form of the varicella-zoster virus (in many countries, Zostavax has now been superseded by the more effective Shingrix vaccine). The virus causes chickenpox in childhood and can reactivate later to cause shingles, a painful rash that can have long-term effects including hearing loss and blindness. Older people and those with weakened immune systems are most at risk.

In 2013, Wales rolled out Zostavax to those aged 70-79 on September 1. The cut-off meant that anyone turning 80 after that date could get jabbed but everyone older, even by one day, never became eligible.

As Geldsetzer puts it, that policy created “these beautiful comparison groups. If you take 1,000 people born one week, and compare them to 1,000 people born a week later, there shouldn’t be anything [systematically] different between them except for their probability of getting the shingles vaccine.”

The two groups in Wales, comprising over 280,000 people, did show a striking difference over the next seven years: the vaccine-eligible group were 20 per cent less likely to receive a dementia diagnosis. Eric Topol, a US researcher who writes the Ground Truths medical blog, surveyed the evidence and concluded recently: “If this vaccine was a drug and reduced Alzheimer’s by 20 per cent, it would be considered a major breakthrough.” One study published last month in The Journals of Gerontology even suggested a link with slower biological ageing.

It is unclear why women benefit more than men and unknown whether the dementia advantage comes from keeping the shingles virus battened down or from a more general vaccine-induced polishing of the immune system. It could be both. Research from 2024 showed that the newer Shingrix vaccine and RSV vaccine are both linked to an even greater reduction of dementia risk, attributed to an immune-boosting ingredient. One theory is that the vaccines somehow reduce inflammation, which might otherwise prime brain cells to go rogue or reactivate viruses. It is unclear, though, how these findings relate to people who have already had shingles, nor whether a jab would lower their dementia risk.

Geldsetzer is now looking to fund a clinical trial on Zostavax to answer those questions and as a step towards regulatory approval. In November, the Alzheimer’s Society, together with scientists at the University of Exeter, also identified Zostavax as a promising candidate.

The UK recently announced plans to lower the age of eligibility for Shingrix to 60; the UK Health Security Agency said vaccine advisers do weigh up off-target benefits and had noted the link. Topol thinks over-50s should consider it. Zostavax, meanwhile, is out of patent. You can’t help but hope that this (relatively) cheap shot — with potentially profound implications for patients, medicine and future dementia research — is on the money.