DETROIT (AP) — Ford is recalling nearly 43,000 small SUVs because gasoline can leak from the fuel injectors onto hot engine surfaces, increasing the risk of fires. But the recall remedy does not include repairing the fuel leaks.

The recall covers certain Bronco Sport SUVs from the 2022 and 2023 model years, as well as Escape SUVs from 2022. All have 1.5-liter engines.

Ford says in documents filed with U.S. safety regulators that fuel injectors can crack, and gasoline or vapor can accumulate near ignition sources, possibly touching off fires.

Dealers will install a tube to let gasoline flow away from hot surfaces to the ground below the vehicle. They’ll also update engine control software to detect a pressure drop in the fuel injection system. If that happens, the software will disable the high pressure fuel pump, reduce engine power and cut temperatures in the engine compartment, according to documents posted Wednesday on the National Highway Traffic Safety Administration website.

Owners were to be notified by letter starting April 1.

The company says in documents it has reports of five under-hood fires and 14 warranty replacements of fuel injectors, but no reports of crashes or injuries.

In an email, Ford said it is not replacing fuel injectors because it is confident the recall repairs “will prevent the failure from occurring and protect the customer.” The new software also will trigger a dashboard warning light and allow customers to drive to a safe location, stop the vehicle and arrange for service, the company said. NHTSA documents filed by Ford say the problem happens only in about 1% of the SUVs.

The company also said it will extend warranty coverage for cracked fuel injectors, so owners who experience the problem will get replacements. Ford said repairs are already available, and details of the extended warranty will be available in June.

Ford said the recall is an extension of a 2022 recall for the same problem. The repair has already been tested on vehicles involved in the previous recall, and Ford said it’s not aware of any problems.

The company also said it isn’t recommending that the SUVs be parked only outdoors because there’s no evidence that fires happen when vehicles are parked and the engines are off.

Michael Brooks, executive director of the nonprofit Center for Auto Safety, called Ford’s remedy for the fuel leaks a “Band-aid type recall” and said the company is trying to avoid the cost of repairing the fuel injectors.

A 1% failure rate, he said, is high, and even with the repairs, drivers still could be forced to exit a freeway at a low speed, placing them at risk of a crash.

NHTSA, he said, should do more to make sure recalls fix the root causes of vehicle problems rather than making less-costly repairs.

In the past, NHTSA has said it does not have legal authority to pre-approve recall fixes. But in a statement Wednesday, the agency said it will “closely track their performance using field data.” The agency said owners who have questions should contact their dealership or Ford.

Brooks said Congress should change the law so the agency can “require something more than the rubber stamp that NHTSA is currently deploying” on recalls.

The agency, he said, has been more aggressive of late in investigating recall fixes. “That is a post-remedy inquiry that won’t make the fixes better, and further stretches out the process and leaves consumers in limbo,” he said.

CLEVELAND (AP) — The women’s NCAA championship game drew a bigger television audience than the men’s title game for the first time, with an average of 18.9 million viewers watching undefeated South Carolina beat Iowa and superstar Caitlin Clark, according to ratings released Tuesday.

The Sunday afternoon game on ABC and ESPN outdrew Monday’s men’s final between UConn and Purdue by four million. The Huskies’ 75-60 victory averaged 14.82 million on TBS and TNT.

The audience for the women’s game — in which the Gamecocks won their fourth national title and denied Clark’s Hawkeyes their first — peaked at 24.1 million during the final 15 minutes.

“You’re seeing the growth in many places: attendance records, viewership and social media engagement surrounding March Madness,” UCLA coach Cori Close said. “I don’t think you can attribute it just to Iowa, though. A rising tide does lifts all boats. But I think all those boats have been on many different waterways. The product is really good, and the increase of exposure is getting rewarded.”

It was the second most-watched non-Olympic women’s sporting event on U.S. television, trailing only the 2015 Women’s World Cup final between the United States and Japan, which averaged 25.4 million on Fox. That also was on a Sunday and took place in prime time on the East Coast.

The record for the most-watched women’s basketball game still belongs to the gold medal game of the 1996 Atlanta Olympics between the United States and Brazil, which averaged 19.5 million. South Carolina coach Dawn Staley played for that U.S. team.

Nielsen’s numbers include an estimate of the number of people who watched outside their homes, which wasn’t measured before 2020. Due to cord-cutting, the in-home audience has steadily declined annually.

The audience for the national title game was up 90% over last year when Clark and Iowa fell to LSU. That also was the first time since 1995 that the championship was on network television.

The audience was 289% bigger than the viewership for the Gamecocks’ title two years ago, when they beat UConn on ESPN.

“I had not seen it much (women’s basketball) before this year. I didn’t make it appointment television. This year, it was appointment television,” said former CBS Sports president Neal Pilson, who now runs a sports television consulting company. “That’s what happened when you see those numbers. There were a lot of people making notes to sit down and watch the games.”

During the Final Four, Clark said the audience growth was benefiting all of women’s sports, not just basketball.

“I think you see it across the board, whether it’s softball, whether it’s gymnastics, volleyball. People want to watch. It’s just when they’re given the opportunity, the research and the facts show that people love it,” she said.

Clark and Iowa have the three biggest audiences for women’s college basketball. The Hawkeyes’ victory over UConn on Friday night averaged 14.2 million, and their April 1 victory over LSU in the Elite Eight drew 12.3 million.

Iowa’s six NCAA Tournament games on ESPN and ABC averaged 10.07 million.

However, as Clark heads to the WNBA, many wonder if the college game can continue to attract large audiences.

Clark was the marquee player of March Madness, but there was an impressive group of freshmen, headlined by Southern California’s JuJu Watkins, Notre Dame’s Hannah Hidalgo and Texas’ Madison Booker. Watkins set the NCAA record for points by a freshman with 920.

“There are also some great freshmen coming into the game next year who I think are going to make a tremendous impact. As long as we keep telling the story, there’s still room to grow the game,” Close said.

It also helped that the women’s game got increased exposure during the regular season on network television.

Fox carried 14 women’s games this season, including three in prime time, NBC two, and CBS had the Big Ten Tournament championship. ABC had five regular-season games and nine during the NCAA Tournament.

Fox has announced the Women’s Champions Classic for next season on Dec. 7. The prime-time doubleheader will feature UConn, Iowa, Louisville and Tennessee.

Pilson thinks this year’s tournament has paved the way for further growth. Just maybe not at the level seen with Clark’s following.

“I wouldn’t call it a blip because I think the women’s game is going to definitely improve from ratings and exposure here on out,” he said. “The broadcast and top cable sports channels can still deliver an audience. Midseason coverage tends to focus on the teams and personalities to make the public aware of the athletes. But the game has improved over the last five years. It has become faster and more competitive. I also think they have learned from the NBA and NFL on how to promote its stars.”

Most audiences during March Madness also tuned in before and after Iowa’s games.

UConn’s 80-73 win over Southern California in the Elite Eight on April 1, which tipped off after the Iowa game, averaged 6.7 million. During the Sweet 16 on March 30, LSU’s victory over UCLA, which preceded the Iowa game, averaged 3.8 million.

“The great thing about what’s happened the last two years is there are people who only watch out for Caitlin Clark, but there are also people who watch for Caitlin Clark and became interested in the other games,” said Jon Lewis, who runs the Sports Media Watch site. “It’s not like the Caitlin Clark games did amazingly well and every other game was at the same level that it was two years ago. They were also tuning in on days when Caitlin Clark didn’t play, which is really notable. Now, none of those numbers are at the level of what we saw for the Iowa games, but they are a lot better than what the tournament was getting before.”

The championship game’s return to network television has also benefited the women’s game. UConn’s victory over Tennessee in 1995 averaged 7.44 million on CBS. Despite ESPN’s work raising the profile of the tournament since it acquired the rights in 1996, the closest the network got to that number for the final was in 2002, when 5.68 million watched UConn beat Oklahoma.

“I think our game has been good for a long time and I think people have just missed the boat. Now we’ve finally had the exposure, and people have understood, ‘Wow, I haven’t watched women’s basketball for a long time, I’ve missed something.’ I don’t think they’re going to want to miss anymore,” Iowa coach Lisa Bluder said during the Final Four. “Caitlin has certainly been a tremendous star for our game, but there are so many stars in our game. So we’re just going to latch onto that next one next year.”

SANTA CLARA, Calif. (AP) — Tesla has settled a lawsuit brought by the family of a Silicon Valley engineer who died in a crash while relying on the company’s semi-autonomous driving software.

The amount Tesla paid to settle the case was not disclosed in court documents filed Monday, just a day before the trial stemming from the 2018 crash on a San Francisco Bay Area highway was scheduled to begin. In a court filing requesting to keep the sum private, Tesla said it agreed to settle the case in order to “end years of litigation.”

The family of Walter Huang filed a negligence and wrongful death lawsuit in 2019 seeking to hold Tesla — and, by extension, its CEO Elon Musk — liable for repeatedly exaggerating the capabilities of Tesla’s self-driving car technology. They claimed the technology, dubbed Autopilot, was promoted in egregious ways that caused vehicle owners to believe they didn’t have to remain vigilant while they were behind the wheel.

Evidence indicated that Huang was playing a video game on his iPhone when he crashed into a concrete highway barrier on March 23, 2018.

After dropping his son off at preschool, Huang activated the Autopilot feature on his Model X for his commute to his job at Apple. But less than 20 minutes later, Autopilot veered the vehicle out of its lane and began to accelerate before barreling into a barrier located at a perilous intersection on a busy highway in Mountain View, California. The Model X was still traveling at more than 70 miles per hour (110 kilometers per hour).

Huang, 38, died at the gruesome scene, leaving behind his wife and two children, now 12 and 9 years old.

The case was just one of about a dozen scattered across the U.S. raising questions about whether Musk’s boasts about the effectiveness of Tesla’s autonomous technology fosters a misguided faith the technology, The company also has an optional feature it calls Full Self Driving. The U.S. Justice Department also opened an inquiry last year into how Tesla and Musk promote its autonomous technology, according to regulatory filings that didn’t provide many details about the nature of the probe.

Tesla, which is based in Austin, Texas, prevailed last year in a Southern California trial focused on whether misperceptions about Tesla’s Autopilot feature contributed to a driver in a 2019 crash involving one of the company’s cars.