ІНСАЙДЕР УКРАЇНА

WASHINGTON — The United States and Britain on Monday announced a new partnership on the science of artificial intelligence safety, amid growing concerns about upcoming next-generation versions.

Commerce Secretary Gina Raimondo and British Technology Secretary Michelle Donelan signed a memorandum of understanding in Washington to jointly develop advanced AI model testing, following commitments announced at an AI Safety Summit in Bletchley Park in November.

“We all know AI is the defining technology of our generation,” Raimondo said. “This partnership will accelerate both of our institutes work across the full spectrum to address the risks of our national security concerns and the concerns of our broader society.”

Britain and the United States are among countries establishing government-led AI safety institutes.

Britain said in October its institute would examine and test new types of AI, while the United States said in November it was launching its own safety institute to evaluate risks from so-called frontier AI models and is now working with 200 companies and entites.

Under the formal partnership, Britain and the United States plan to perform at least one joint testing exercise on a publicly accessible model and are considering exploring personnel exchanges between the institutes. Both are working to develop similar partnerships with other countries to promote AI safety.

“This is the first agreement of its kind anywhere in the world,” Donelan said. “AI is already an extraordinary force for good in our society and has vast potential to tackle some of the world’s biggest challenges, but only if we are able to grip those risks.”

Generative AI, which can create text, photos and videos in response to open-ended prompts, has spurred excitement as well as fears it could make some jobs obsolete, upend elections and potentially overpower humans and catastrophic effects.

In a joint interview with Reuters Monday, Raimondo and Donelan urgent joint action was needed to address AI risks.

“Time is of the essence because the next set of models are about to be released, which will be much, much more capable,” Donelan said. “We have a focus one the areas that we are dividing and conquering and really specializing.”

Raimondo said she would raise AI issues at a meeting of the U.S.-EU Trade and Technology Council in Belgium Thursday.

The Biden administration plans to soon announce additions to its AI team, Raimondo said. “We are pulling in the full resources of the U.S. government.”

Both countries plan to share key information on capabilities and risks associated with AI models and systems and technical research on AI safety and security.

In October, Biden signed an executive order that aims to reduce the risks of AI. In January, the Commerce Department said it was proposing to require U.S. cloud companies to determine whether foreign entities are accessing U.S. data centers to train AI models.

Britain said in February it would spend more than 100 million pounds ($125.5 million) to launch nine new research hubs and AI train regulators about the technology.

Raimondo said she was especially concerned about the threat of AI applied to bioterrorism or a nuclear war simulation.

“Those are the things where the consequences could be catastrophic and so we really have to have zero tolerance for some of these models being used for that capability,” she said.
…

…

…

New York — Kia is recalling more than 427,000 of its Telluride SUVs due to a defect that may cause the cars to roll away while they’re parked.

According to documents published by the National Highway Traffic Safety Administration, the intermediate shaft and right front driveshaft of certain 2020-2024 Tellurides may not be fully engaged. Over time, this can lead to “unintended vehicle movement” while the cars are in park — increasing potential crash risks.

Kia America decided to recall all 2020-2023 model year and select 2024 model year Tellurides earlier this month, NHTSA documents show. At the time, no injuries or crashes were reported.

Improper assembly is suspected to be the cause of the shaft engagement problem — with the recall covering 2020-2024 Tellurides that were manufactured between Jan. 9, 2019, and Oct. 19, 2023. Kia America estimates that 1% have the defect.

To remedy this issue, recall documents say, dealers will update the affected cars’ electronic parking brake software and replace any damaged intermediate shafts for free. Owners who already incurred repair expenses will also be reimbursed.

In the meantime, drivers of the impacted Tellurides are instructed to manually engage the emergency brake before exiting the vehicle. Drivers can also confirm if their specific vehicle is included in this recall and find more information using the NHTSA site and/or Kia’s recall lookup platform.

Owner notification letters are otherwise set to be mailed out on May 15, with dealer notification beginning a few days prior.

The Associated Press reached out to Irvine, California-based Kia America for further comment Sunday. No comment was received.
…

San Francisco — Google co-founders Larry Page and Sergey Brin loved pulling pranks, so they began rolling out outlandish ideas every April Fool’s Day not long after starting their company more than a quarter century ago. One year, Google posted a job opening for a Copernicus research center on the moon. Another year, the company said it planned to roll out a “scratch and sniff” feature on its search engine.

The jokes were consistently over-the-top, and people learned to laugh them off as another example of Google mischief. That’s why Page and Brin decided to unveil something no one would believe was possible 20 years ago on April Fool’s Day.

It was Gmail, a free service boasting 1 gigabyte of storage per account, an amount that sounds almost pedestrian in an age of 1-terabyte iPhones. But it sounded like a preposterous amount of email capacity back then, enough to store about 13,500 emails before running out of space compared to just 30 to 60 emails in the then-leading webmail services run by Yahoo and Microsoft. That translated into 250 to 500 times more email storage space.

Besides the quantum leap in storage, Gmail also came equipped with Google’s search technology so users could quickly retrieve a tidbit from an old email, photo or other personal information stored on the service. It also automatically threaded together a string of communications about the same topic, so everything flowed together as if it was a single conversation.

“The original pitch we put together was all about the three ‘S’s’ — storage, search and speed,” said former Google executive Marissa Mayer, who helped design Gmail and other company products before later becoming Yahoo’s CEO.

It was such a mind-bending concept that shortly after The Associated Press published a story about Gmail late on the afternoon of April Fool’s 2004, readers began calling and emailing to inform the news agency it had been duped by Google’s pranksters.

“That was part of the charm, making a product that people won’t believe is real. It kind of changed people’s perceptions about the kinds of applications that were possible within a web browser,” former Google engineer Paul Buchheit recalled during a recent AP interview about his efforts to build Gmail.

It took three years to do as part of a project called “Caribou” — a reference to a running gag in the Dilbert comic strip. “There was something sort of absurd about the name Caribou, it just made make me laugh,” said Buchheit, the 23rd employee hired at a company that now employs more than 180,000 people.

The AP knew Google wasn’t joking about Gmail because an AP reporter had been abruptly asked to come down from San Francisco to the company’s Mountain View, California, headquarters to see something that would make the trip worthwhile.

After arriving at a still-developing corporate campus that would soon blossom into what became known as the “Googleplex,” the AP reporter was ushered into a small office where Page was wearing an impish grin while sitting in front of his laptop computer.

Page, then just 31 years old, proceeded to show off Gmail’s sleekly designed inbox and demonstrated how quickly it operated within Microsoft’s now-retired Explorer web browser. And he pointed out there was no delete button featured in the main control window because it wouldn’t be necessary, given Gmail had so much storage and could be so easily searched. “I think people are really going to like this,” Page predicted.

As with so many other things, Page was right. Gmail now has an estimated 1.8 billion active accounts — each one now offering 15 gigabytes of free storage bundled with Google Photos and Google Drive. Even though that’s 15 times more storage than Gmail initially offered, it’s still not enough for many users who rarely see the need to purge their accounts, just as Google hoped.

The digital hoarding of email, photos and other content is why Google, Apple and other companies now make money from selling additional storage capacity in their data centers. (In Google’s case, it charges anywhere from $30 annually for 200 gigabytes of storage to $250 annually for 5 terabytes of storage). Gmail’s existence is also why other free email services and the internal email accounts that employees use on their jobs offer far more storage than was fathomed 20 years ago.

“We were trying to shift the way people had been thinking because people were working in this model of storage scarcity for so long that deleting became a default action,” Buchheit said.

Gmail was a game changer in several other ways while becoming the first building block in the expansion of Google’s internet empire beyond its still-dominant search engine.

After Gmail came Google Maps and Google Docs with word processing and spreadsheet applications. Then came the acquisition of video site YouTube, followed by the introduction of the Chrome browser and the Android operating system that powers most of the world’s smartphones. With Gmail’s explicitly stated intention to scan the content of emails to get a better understanding of users’ interests, Google also left little doubt that digital surveillance in pursuit of selling more ads would be part of its expanding ambitions.

Although it immediately generated a buzz, Gmail started out with a limited scope because Google initially only had enough computing capacity to support a small audience of users.

But that scarcity created an air of exclusivity around Gmail that drove feverish demand for elusive invitations to sign up. At one point, invitations to open a Gmail account were selling for $250 apiece on eBay. “It became a bit like a social currency, where people would go, ‘Hey, I got a Gmail invite, you want one?’” Buchheit said.

Although signing up for Gmail became increasingly easier as more of Google’s network of massive data centers came online, the company didn’t begin accepting all comers to the email service until it opened the floodgates as a Valentine’s Day present to the world in 2007.
…

WASHINGTON — Artificial Intelligence for good is the subject of a new exhibit at the Embassy of Sweden in Washington, showing how Swedish companies and organizations are using AI for a more open society, a healthier world, and a greener planet.

Ambassador Urban Ahlin told an embassy reception that Sweden’s broad collaboration across industry, academia and government makes it a leader in applying AI in public-interest areas, such as clean tech, social sciences, medical research, and greener food supply chains. That includes tracking the mood and health of cows.

Fitbit for cows

It is technology developed by DeLaval, a producer of dairy and farming machinery. The firm’s Market Solution Manager in North America Joaquin Azocar says the small wearable device the size of an earring fits in a cow’s ear and tracks the animal’s movements 24/7, much like a Fitbit.

The ear-mounted tags send out signals to receivers across the farm. DeLaval’s artificial intelligence system analyzes the data and looks for correlations in patterns, trends, and deviations in the animals’ activities, to predict if a cow is sick, in heat, or not eating well.

As a trained veterinarian, Azocar says dairy farmers being alerted sooner to changes in their animals’ behavior means they can provide treatment earlier which translates to less recovery time.

AI helping in childbirth

There are also advances in human health. The developing Pelvic Floor AI project is an AI-based solution to identify high-risk cases of pelvic floor injury and facilitate timely interventions to prevent and limit harm.

It was developed by a team of gynecologists and women’s health care professionals from Sweden’s Sahlgrenska University Hospital to help the nearly 20% of women who experience injury to their pelvic floor during childbirth.

The exhibition “is a great way to showcase the many ways AI is being adapted and used, in medicine and in many other areas,” said exhibition attendee Jesica Lindgren, general counsel for international consulting firm BlueStar Strategies. “It’s important to know how AI is evolving and affecting our everyday life.”

Green solutions using AI

The exhibition includes examples of what AI can do about climate change, including rising sea levels and declining biodiversity.

AirForestry is developing technology “for precise forestry that will select and harvest trees fully autonomously.” The firm says that “harvesting the right trees in the right place could significantly improve overall carbon sequestration and resilience.”

AI & the defense industry

Outlining the development of artificial intelligence for the defense industry, the exhibit admits that “can be controversial.”

“There are exciting possibilities to use AI to solve problems that cannot be solved using traditional algorithms due to their complexity and limitations in computational power,” the exhibit states. “But it requires thorough consideration of how AI should and shouldn’t be utilized. Proactively engaging in AI research is necessary to understand the technology’s capabilities and limitations and help shape its ethical standards.”

AI and privacy

Exhibition participant Quentin Black is an engineer with Axis Communications, an industry leader in video surveillance. He said the project came out of GDPR, or General Data Protection Regulation; an EU policy that provides privacy to citizens who are out in public whose image could be picked up on video surveillance cameras.

The regulations surrounding privacy are stricter in Europe than they are in the U.S., Black said.

“In the U.S. the public doesn’t really have an expectation of privacy; there’s cameras everywhere. In Europe, it’s different.” That regulation inspired Axis Communications to develop AI that provides privacy, he explained.

Black pointed to a large monitor divided into four windows, to show how AI is being used to set up four different filters to provide privacy.

The Axis Live Privacy Shield remotely monitors activities both indoors and outdoors while safeguarding privacy in real time. The technology is downloadable and free, to provide privacy to people and/or environments, using a variety of filters.

In the monitor on display in the exhibition, Black explained the four quadrants. The upper right window of the monitor displays privacy with a full color block out of all humans, using AI to distinguish the difference between the people and the environment.

The upper left window provides privacy to the person’s head. The bottom left corner provides pixelization, or a mosaic, of the person’s entire/whole body, and the immediate environment surrounding the person. And the bottom right corner shows blockage of the environment, so “an inverse of the personal privacy,” Black explained.

“So, if it was a top secret facility, or you want to see the people walking up to your door without a view of your neighbor’s house, this is where this can this be applied.”

Tip of the iceberg

“I think that AI is on everybody’s thoughts, and what I appreciate about the House of Sweden’s approach in this exhibition is highlighting a thoughtful, scientific, business-oriented and human-oriented perspective on AI in society today,” said Molly Steenson, President and CEO of the American Swedish Institute.

Though AI and machine learning have been around since the 1950s, she says it is only now that we are seeing “the contemporary upswing and acceleration of AI, especially generative AI in things like large language models.”

“So, while large companies and tech companies might want us to speed up and believe that it is only scary or it is only good, I think it’s a lot more nuanced than that,” she said.
…

A space launch aborted only to find success days later. Plus, Japan makes a push into private spaceflight, and NASA really wants you to see the solar eclipse — but safety first. VOA’s Arash Arabasadi brings us The Week in Space.
…

…

WASHINGTON — The young voices in the messages left for North Carolina Senator Thom Tillis were laughing, but the words were ominous.

“OK, listen, if you ban TikTok I will find you and shoot you,” one said, giggling and talking over other young voices in the background. “I’ll shoot you and find you and cut you into pieces.” Another threatened to kill Tillis, and then take their own life.

Tillis’s office says it has received around 1,000 calls about TikTok since the House passed legislation this month that would ban the popular app if its China-based owner doesn’t sell its stake. TikTok has been urging its users — many of whom are young — to call their representatives, even providing an easy link to the phone numbers. “The government will take away the community that you and millions of other Americans love,” read one pop-up message from the company when users opened the app.

Tillis, who supports the House bill, reported the call to the police. “What I hated about that was it demonstrates the enormous influence social media platforms have on young people,” he said in an interview.

While more aggressive than most, TikTok’s extensive lobbying campaign is the latest attempt by the tech industry to head off any new legislation — and it’s a fight the industry usually wins. For years Congress has failed to act on bills that would protect users’ privacy, protect children from online threats, make companies more liable for their content and put loose guardrails around artificial intelligence, among other things.

“I mean, it’s almost embarrassing,” says Senate Intelligence Committee Chairman Mark Warner, D-Va., a former tech executive who is also supporting the TikTok bill and has long tried to push his colleagues to regulate the industry. “I would hate for us to maintain our perfect zero batting average on tech legislation.”

Some see the TikTok bill as the best chance for now to regulate the tech industry and set a precedent, if a narrow one focused on just one company. President Joe Biden has said he would sign the House bill, which overwhelmingly passed 362-65 this month after a rare 50-0 committee vote moving it to the floor.

But it’s already running into roadblocks in the Senate, where there is little unanimity on the best approach to ensure that China doesn’t access private data from the app’s 170 million U.S. users or influence them through its algorithms.

Other factors are holding the Senate back. The tech industry is broad and falls under the jurisdiction of several different committees. Plus, the issues at play don’t fall cleanly on partisan lines, making it harder for lawmakers to agree on priorities and how legislation should be written. Senate Commerce Committee Chairwoman Maria Cantwell, D-Wash., has so far been reluctant to embrace the TikTok bill, for example, calling for hearings first and suggesting that the Senate may want to rewrite it.

“We’re going through a process,” Cantwell said. “It’s important to get it right.”

Warner, on the other hand, says the House bill is the best chance to get something done after years of inaction. And he says that the threatening calls from young people are a good example of why the legislation is needed: “It makes the point, do we really want that kind of messaging being able to be manipulated by the Communist Party of China?”

Some lawmakers are worried that blocking TikTok could anger millions of young people who use the app, a crucial segment of voters in November’s election. But Warner says “the debate has shifted” from talk of an outright ban a year ago to the House bill which would force TikTok, a wholly owned subsidiary of Chinese technology firm ByteDance Ltd., to sell its stake for the app to continue operating.

Vice President Kamala Harris, in a television interview that aired Sunday, acknowledged the popularity of the app and that it has become an income stream for many people. She said the administration does not intend to ban TikTok but instead deal with its ownership. “We understand its purpose and its utility and the enjoyment that it gives a lot of folks,” Harris told ABC’s ”This Week.”

Republicans are divided. While most of them support the TikTok legislation, others are wary of overregulation and the government targeting one specific entity.

“The passage of the House TikTok ban is not just a misguided overreach; it’s a draconian measure that stifles free expression, tramples constitutional rights, and disrupts the economic pursuits of millions of Americans,” Kentucky Sen. Rand Paul posted on X, formerly Twitter.

Hoping to persuade their colleagues to support the bill, Democratic Sen. Richard Blumenthal of Connecticut and Republican Sen. Marsha Blackburn of Tennessee have called for intelligence agencies to declassify information about TikTok and China’s ownership that has been provided to senators in classified briefings.

“It is critically important that the American people, especially TikTok users, understand the national security issues at stake,” the senators said in a joint statement.

Blumenthal and Blackburn have separate legislation they have been working on for several years aimed at protecting children’s online safety, but the Senate has yet to vote on it. Efforts to regulate online privacy have also stalled, as has legislation to make technology companies more liable for the content they publish.

And an effort by Senate Majority Leader Chuck Schumer, D-N.Y., to quickly move legislation that would regulate the burgeoning artificial intelligence industry has yet to show any results.

Schumer has said very little about the TikTok bill or whether he might put it on the Senate floor.

“The Senate will review the legislation when it comes over from the House,” was all he would say after the House passed the bill.

South Dakota Sen. Mike Rounds, a Republican who has worked with Schumer on the artificial intelligence effort, says he thinks the Senate can eventually pass a TikTok bill, even if it’s a different version. He says the classified briefings “convinced the vast majority of members” that they have to address the collection of data from the app and TikTok’s ability to push out misinformation to users.

“I think it’s a clear danger to our country if we don’t act,” he said. “It does not have to be done in two weeks, but it does have to be done.”

Rounds says he and Schumer are still holding regular meetings on artificial intelligence, as well, and will soon release some of their ideas publicly. He says he’s optimistic that the Senate will eventually act to regulate the tech industry.

“There will be some areas that we will not try to get into, but there are some areas that we have very broad consensus on,” Rounds says.

Tillis says senators may have to continue laying the groundwork for a while and educating colleagues on why some regulation is needed, with an eye toward passing legislation in the next Congress.

“It can’t be the wild, wild west,” Tillis said.
…