ChatGPT could return to Italy soon if its maker, OpenAI, complies with measures to satisfy regulators who had imposed a temporary ban on the artificial intelligence software over privacy worries.

The Italian data protection authority on Wednesday outlined a raft of requirements that OpenAI will have to satisfy by April 30 for the ban on AI chatbot to be lifted.

The watchdog last month ordered the company to temporarily stop processing Italian users’ personal information while it investigated a possible data breach. The authority said it didn’t want to hamper AI’s development but emphasized the importance of following the European Union’s strict data privacy rules.

OpenAI, which had responded by proposing remedies to ease the concerns, did not reply immediately to a request for comment Wednesday.

Concerns about boom grow

Concerns are growing about the artificial intelligence boom, with other countries, from France to Canada, investigating or looking closer at so-called generative AI technology like ChatGPT. The chatbot is “trained” on huge pools of data, including digital books and online writings, and able to generate text that mimics human writing styles.

Under Italy’s measures, OpenAI must post information on its website about how and why it processes the personal information of both users and non-users, as well as provide the option to correct or delete that data.

The company will have to rely on consent or “legitimate interest” to use personal data to train ChatGPT’s algorithms, the watchdog said.

Regulators question legal basis

The Italian regulators had questioned whether there’s a legal basis for OpenAI to collect massive amounts of data used to teach ChatGPT’s algorithms and raised concerns the system could sometimes generate false information about individuals.

San Francisco-based OpenAI also will have to carry out a publicity campaign by May 15 through radio and TV, newspapers and the internet to inform people about how it uses their personal data for training algorithms, Italy’s watchdog said.

There’s also a requirement to verify users’ ages and set up a system to filter out those who are under 13 and teens between 13 and 18 who don’t have parental consent.

“Only in that case will the Italian SA (supervisory authority) lift its order that placed a temporary limitation on the processing of Italian users’ data … so that ChatGPT will be available once again from Italy,” the watchdog said on its website.
…

A Biden administration proposal would force U.S. automakers to sharply increase their production of electric cars and trucks over the next decade, lending greater urgency to the effort to build raw material supply chains that reduce the industry’s dependence on China.

The Environmental Protection Agency on Wednesday announced a proposed rule that would place stricter limits on the average tailpipe emissions of vehicles built in the United States. The proposal would reduce the allowable limit by so much that automakers would have no way to comply unless about two-thirds of the vehicles they produce by 2032 are emission-free electric vehicles.

Automakers have generally recognized that EVs represent the future of the industry, but Wednesday’s proposal would greatly accelerate the trend. The proposal, which will be open to public comment before it is finalized, would greatly reduce a leading cause of air pollution in the U.S., as well as the greenhouse gases that contribute to global warming.

“By proposing the most ambitious pollution standards ever for cars and trucks, we are delivering on the Biden-Harris administration’s promise to protect people and the planet, securing critical reductions in dangerous air and climate pollution, and ensuring significant economic benefits like lower fuel and maintenance costs for families,” said EPA Administrator Michael Regan.

The proposal, which would apply to new light-duty vehicles made in 2027 and beyond, would be the strictest environmental standard the federal government has ever applied to automobiles. If it does force the industry to make EVs account for two-thirds of production, it could also exceed President Joe Biden’s previously articulated target of making 50% of new cars either plug-in hybrids or completely emission-free by 2030.

Supply chain questions

Well before the EPA released its proposed rule Wednesday, the Biden administration had been moving to strengthen the EV market in the U.S. and to build a pipeline for raw materials that would reduce the auto industry’s reliance on China for key raw materials.

Accomplishing that reduction will be no small task. According to an analysis by the International Energy Agency last year, China produced three-quarters of the world’s lithium-ion batteries, the key component in the majority of EVs on the road.

China also has a dominant hold on much of the market for the components of those batteries, including lithium, cobalt and graphite. According to the IEA, more than half of the world’s capacity for processing and refining those materials is located in China.

According to the IEA, as of last year, the U.S. accounted for only 10% of EV production worldwide, and just 7% of production capacity for batteries.

Infrastructure projects

Last year’s passage of the Inflation Reduction Act, which contained hundreds of billions of dollars in climate-related spending, included the creation of large tax breaks restricted to EVs made at least partly in the U.S. The tax breaks are meant to extend over several years, but the restrictions become tighter as time goes on, creating incentives for manufacturers to “onshore” production to the U.S.

Tax breaks specific to the batteries used in EVs require that the raw materials used to assemble them come from domestic sources or from countries with which the U.S. has existing trade agreements.

Other pieces of legislation meant to spur investment in the U.S., including a major bipartisan infrastructure bill and the CHIPS and Science Act, also contain money and incentives that will help build out electric infrastructure in the U.S.

Achievable goals

Luke Tonachel, senior director for clean vehicles and buildings with the Natural Resources Defense Council, told VOA that building an EV supply chain centered on domestic production and imports from friendly countries is ambitious, but achievable.

Tonachel said the necessary raw materials are available from U.S. allies, but that the capacity for processing them needs to be built domestically. He said the creation of that capacity is already underway.

“There are robust incentives for building out that battery manufacturing and supply chain here in the U.S.,” he said, adding that he believes the administration’s time frame is feasible, especially now that the new standards have created certainty about future demand for EVs.

“It is realistic,” he said. “These are technologies that are known. We can certainly get more economies of scale as we ramp up production.”

Automakers tentative

Industry representatives said achieving the administration’s goal will require that a lot of disparate efforts be successful at the same time, not all of which are under their control. For example, a nationwide network of charging stations and the increased capacity to meet new demand for power will be essential to driving customer demand.

“It’s aggressive, and a lot of pieces have to work perfectly together,” Genevieve Cullen, president of the Electric Drive Transportation Association, told VOA. “Aside from the technology piece, the market piece has to work, and supply chain speed is part of that. Consumer incentives are working to help bring them into the equation, and we need to keep expanding infrastructure at a pace that meets, and perhaps exceeds, the needs in the beginning so that people feel the confidence that they need to switch to battery electric.”

John Bozzella, president of the trade group Alliance for Automotive Innovation, said in a blog post Wednesday that the administration’s plan is “aggressive by any measure” and that its success would depend on more than just automakers being able to ramp up production.

“To some extent, the baseline policy framework for the transition has come into focus,” Bozzella said. “But it remains to be seen whether the refueling infrastructure incentives and supply-side provisions of the Inflation Reduction Act, the bipartisan infrastructure law, and the CHIPS and Science Act are sufficient to support electrification at the levels envisioned by the proposed standards over the coming years.”
…

Broadcaster National Public Radio said Wednesday it would no longer post its news content on 52 official Twitter accounts in protest of the social media site labeling the independent U.S. news agency as “government-funded media.” 

NPR is the first major news organization to go silent on Twitter. The social media platform owned by entrepreneur Elon Musk at first labeled NPR as “state-affiliated media,” the same tag it applies to propaganda outlets in China, Russia and other autocratic countries. 

Twitter then revised its label to “government-funded media,” but NPR said that, too, was misleading because NPR is a private, nonprofit company with editorial independence. NPR says it receives less than 1% of its $300 million annual budget from the federally funded Corporation for Public Broadcasting.  

NPR chief executive John Lansing said that by not posting its news reports on Twitter, the network is protecting its credibility and would continue to produce journalism without “a shadow of negativity.” 

In an email to staff explaining the decision, Lansing wrote, “It would be a disservice to the serious work you all do here to continue to share it on a platform that is associating the federal charter for public media with an abandoning of editorial independence or standards.”  

He said that even if Twitter were to drop any description of NPR, the network would not immediately return to the platform. 

“At this point I have lost my faith in the decision-making at Twitter,” Lansing said in an article posted by NPR. “I would need some time to understand whether Twitter can be trusted again.”

Twitter has also labeled Voice of America, a U.S. government-funded but independent news agency, and the BBC in Britain, as “government-funded media,” a description more commonly employed in describing state-controlled propaganda outlets. VOA has not dropped its use of Twitter but said its description of the news outlet left the impression that it was not independent. 

Bridget Serchak, VOA’s director of public relations, said, “The label ‘government funded’ is potentially misleading and could be construed as also ‘government-controlled’ — which VOA is most certainly not.” 

“Our editorial firewall, enshrined in the law, prohibits any interference from government officials at any level in its news coverage and editorial decision-making process,” Serchak said in an email. “VOA will continue to emphasize this distinction in our discussions with Twitter, as this new label on our network causes unwarranted and unjustified concern about the accuracy and objectivity of our news coverage.” 

VOA is funded by the U.S. government and is part of the U.S. Agency for Global Media, but its editorial independence is protected by regulations and a firewall. The BBC said it “is, and always has been, independent.” 

Press freedom advocates have also objected to Twitter’s labeling of NPR, VOA and the BBC.

“The confusion between media serving the general interest and propaganda media is dangerous, and is yet further proof that social media platforms are not competent to identify what is and is not journalism,” Vincent Berthier, head of the technology desk at Reporters Without Borders, said in a statement. 

Liam Scott contributed to this report.
…

Billionaire Elon Musk has told the BBC that running Twitter has been “quite painful” but that the social media company is now roughly breaking even after he acquired it late last year.

In an interview also streamed live late Tuesday on Twitter Spaces, Musk discussed his ownership of the online platform, including layoffs, misinformation and his work style.

“It’s not been boring. It’s quite a rollercoaster,” he told the U.K. broadcaster at Twitter’s San Francisco headquarters.

It was a rare chance for a mainstream news outlet to interview Musk, who also owns Tesla and SpaceX. After buying Twitter for $44 billion last year, Musk’s changes included eliminating the company’s communications department.

Reporters who email the company to seek comment now receive an auto-reply with a poop emoji.

The interview was sometimes tense, with Musk challenging the reporter to back up assertions about rising levels of hate speech on the platform. At other times, Musk laughed at his own jokes, mentioning more than once that he wasn’t the CEO but his dog Floki was.

He also revealed that he sometimes sleeps on a couch at Twitter’s San Francisco office.

Advertisers who had shunned the platform in the wake of Musk’s tumultuous acquisition have mostly returned, the billionaire said, without providing details.

Musk predicted that Twitter could become “cash flow positive” in the current quarter “if current trends continue.” Because Twitter is a private company, information about its finances can’t be verified.

After acquiring the platform, Musk carried out mass layoffs as part of cost-cutting efforts. He said Twitter’s workforce has been slashed to about 1,500 employees from about 8,000 previously, describing it as something that had to be done.

“It’s not fun at all,” Musk said. “The company’s going to go bankrupt if we don’t cut costs immediately. This is not a caring-uncaring situation. It’s like if the whole ship sinks, then nobody’s got a job.”

Asked if he regretted buying the company, he said it was something that “needed to be done.”

“The pain level of Twitter has been extremely high. This hasn’t been some sort of party,” Musk said.
…

China’s internet regulator has unveiled a proposed law that will require makers of new artificial intelligence, or AI, products to submit to security assessments before public release.

The draft law released Tuesday by the Cyberspace Administration of China says that content generated by future AI products must reflect the country’s “core socialist values” and not encourage subversion of state power.  

The draft law also said AI content must not promote discrimination based on ethnicity, race and gender, and should not provide false information.  

The proposed law is expected to take effect sometime this year. The regulations come as several China-based tech companies, including Alibaba, JD.com and Baidu have released a flurry of new so-called generative AI products which can mimic human speech and generate content such as images and texts. The innovative feature has surged in popularity since San Francisco-based OpenAI introduced ChatGPT last November.  

Some information for this report came from Reuters, Agence France-Presse.
…

The Australian government is asking major banks and other institutions to take part in ‘wargaming’ exercises to test how they would respond to cyber-attacks. It follows recent mass data theft attacks on several large companies, which compromised the data of millions of Australians.     

Australia is preparing for potential cyberattacks on critical services including hospitals, the banking system and the electricity grid.  

Home Affairs Minister Clare O’Neil Tuesday warned that recent high-profile hacks on the telecommunications and health insurance sectors, which have affected millions of people, “were just the tip of the iceberg”.   

The government is setting up a series of drills with large organizations to help them respond to security breaches.   

Anna Bligh, chief executive of the Australian Banking Association, an industry body, told the Australian Broadcasting Corp. Tuesday that cyber security drills organized by the government will make the sector more resilient. 

“How would the whole system cope if one of the very large companies were taken down by a cyber threat?” Bligh asked. “The sort of scale and sophistication of the threat is now moving into something that we haven’t seen before. So, it is a very timely move.  This is now potentially a significant threat to the national security of the country.”   

A major Australian financial services company revealed Tuesday that criminals who stole sensitive customer information last month have demanded a ransom.   

The cyberattack on Latitude Financial resulted in the theft of 14 million customer records, including financial statements, driver’s license numbers and passport numbers.   

The company said that in line with government policy it would not pay a ransom to prevent the data being leaked or sold online.   

The Australian government is considering an updated Cyber Security law that would impose new obligations and standards to protect data across industry and government departments.   

However, officials have warned that cyber criminals are becoming more professional, powerful and effective. 
…

A Kuwaiti media outlet has unveiled a virtual news presenter generated using artificial intelligence, with plans for it to read online bulletins.   

“Fedha” appeared on the Twitter account of the Kuwait News website Saturday as an image of a woman, her light-colored hair uncovered, wearing a black jacket and white T-shirt.   

“I’m Fedha, the first presenter in Kuwait who works with artificial intelligence at Kuwait News. What kind of news do you prefer? Let’s hear your opinions,” she said in classical Arabic.   

The site is affiliated with the Kuwait Times, founded in 1961 as the Gulf region’s first English-language daily.   

Abdullah Boftain, deputy editor-in-chief for both outlets, said the move is a test of AI’s potential to offer “new and innovative content.”   

In the future Fedha could adopt the Kuwaiti accent and present news bulletins on the site’s Twitter account, which has 1.2 million followers, he said.   

“Fedha is a popular, old Kuwaiti name that refers to silver, the metal. We always imagine robots to be silver and metallic in color, so we combined the two,” Boftain said.    

The presenter’s blonde hair and light-colored eyes reflect the oil-rich country’s diverse population of Kuwaitis and expatriates, according to Boftain.    

“Fedha represents everyone,” he said.   

Her initial 13-second video generated a flood of reactions on social media, including from journalists. 

The rapid rise of AI globally has raised the promise of benefits, such as in health care and the elimination of mundane tasks, but also fears, for example over its potential to spread disinformation, threats to certain jobs and to artistic integrity.   

Kuwait ranked 158 out of 180 countries and territories in the Reporters Without Borders (RSF) 2022 Press Freedom Index. 
…

Electric car maker Tesla Inc. plans to build a factory in Shanghai to produce power-storage devices for sale worldwide, state media reported Sunday.

Plans call for annual production of 10,000 Megapack units, according to the Xinhua News Agency and state television. They said the company made the announcement at a signing ceremony in Shanghai, where Tesla operates an auto factory.

The factory is due to break ground in the third quarter of this year and start production in the second quarter of 2024, the reports said.

Tesla didn’t immediately respond to requests for information.
…

A mayor in Australia’s Victoria state said Friday he may sue the artificial intelligence writing tool ChatGPT after it falsely claimed he’d served time in prison for bribery.  Hepburn Shire Council Mayor Brian Hood was incorrectly identified as the guilty party in a corruption case in the early 2000s.

Brian Hood was the whistleblower in a corruption scandal involving a company partly owned by the Reserve Bank of Australia.  Several people were charged, but Hood was not one of them. That did not stop an article generated by ChatGPT, an automated writing service powered by artificial intelligence. The article cast him as the culprit who was jailed for his part in a conspiracy to bribe foreign officials to win currency printing contracts.

Hood only found out after friends told him. He told the Australian Broadcasting Corp. He then used the chatbot software to see the story for himself.

“After making the inquiry, it generated five or six paragraphs of information.  The really disturbing thing was that some of the paragraphs were accurate, and then were other paragraphs that described things that were completely incorrect.  It told me that I’d be charged with very serious criminal offenses, that I’d be convicted of them and that I had spent 30 months in jail,” he said.

Hood said that if OpenAI, a U.S.-based company that owns the chatbot, does not correct the false claims, he will sue.

It would be the first defamation lawsuit against the automated service.

However, a new version of ChatGPT reportedly avoids the mistakes of its predecessor. It reportedly correctly explains that Hood was a whistleblower who was praised for his actions. Hood’s lawyers say that the defamatory material, which damages the mayor’s reputation, still exists and their efforts to have the mistakes rectified would continue.

A disclaimer on the ChatGPT program warns users that it “may produce inaccurate information about people, places, or facts”.  The technology has exploded in popularity around the world.

OpenAI has yet to comment publicly on the allegations.

Google has announced the launch of its rival to ChatGPT, Bard.  Meta, which owns WhatsApp, Facebook and Instagram, launched its own AI chatbot Blenderbot in the United States last year, while Baidu, the Chinese tech company, has said it was working on an advanced version of its chatbot, Ernie.
…

Samsung Electronics said Friday it is cutting the production of its computer memory chips in an apparent effort to reduce inventory as it forecasted another quarter of sluggish profit. 

The South Korean technology giant, in a regulatory filing, said it has been reducing the production of certain memory products by unspecified “meaningful levels” to optimize its manufacturing operations, adding it has sufficient supplies of those chips to meet demand fluctuations. 

The company predicted an operating profit of $455 million for the three months through March, which would be a 96% decline from the same period a year earlier. It said sales during the quarter likely fell 19% to $47.7 billion. 

Samsung, which will release its finalized first quarter earnings later this month, said the demand for its memory chips declined as a weak global economy depressed consumer spending on technology products and forced business clients to adjust their inventories to nurse worsening finances. 

Samsung had reported a near 70% drop in profit for the October-December quarter, which partially reflected how global events like Russia’s war on Ukraine and high inflation have rattled technology markets. 

SK Hynix, another major South Korean semiconductor producer, said this week that it sold $1.7 billion in bonds that can be exchanged for the company’s shares to help fund its purchases of chipmaking materials as it weathers the industry’s downswing. SK Hynix had reported an operating loss of $1.28 billion for the October-December period, which marked its first quarterly deficit since 2012. 

“While we have lowered our short-term production plans, we expect solid demand for the mid- to long-term, so we will continue to invest in infrastructure to secure essential levels in clean room capacities and expand investment in research and development to strengthen our technology leadership,” Samsung said. 

Samsung last month announced plans to invest $227 billion over the next 20 years as part of an ambitious South Korean project to build the world’s largest semiconductor manufacturing base near the capital, Seoul. 

The chip-making “mega cluster,” which will be established in Gyeonggi province by 2042, will be anchored by five new semiconductor plants built by Samsung near its existing manufacturing hub. It will aim to attract 150 other companies producing materials and components or designing high-tech chips, according to South Korea’s government. 

The South Korean plan comes as other technology powerhouses, including the United States, Japan and China, are building up their domestic chip manufacturing, deploying protectionist measures, tax cuts and sizeable subsidies to lure investments. 
…

The last time humans were on the moon was in 1972. Now NASA is preparing to set foot back on the moon in 2025, if all goes as scheduled. VOA’s Alexander Kruglyakov spoke with the crew that will take part in the first of those missions: a planned flight around the Moon in November 2024.
…

Darknet users, beware: If you frequent criminal marketplaces in the internet’s underbelly, think again. Chances are you’re in the FBI’s crosshairs. 

The FBI is cracking down on sites that peddle everything from guns to stolen personal data, and it is not only going after the sites’ administrators but also their users.  

A recent surge in ransomware attacks and other malicious cyber activities has fueled the effort to shut down services that cater to online criminals.  

But shutting down the marketplaces has proven ineffective. With each takedown, a new iteration pops up drawing users with it. Which is why the FBI is eyeing both the operators and users of these sites.   

“We’re not only trying to attack the supply side, but we’re also attacking the demand side with the users,” a senior FBI official said during a Wednesday briefing on the agency’s takedown of Genesis Market, a large online criminal marketplace. “There’s consequences if you’re going to be using these types of sites to engage in this type of activity.” 

The darknet, the hidden part of the internet that can only be accessed by a special browser, has long been home to various criminal marketplaces and forums. 

One type of criminal marketplace there specializes in buying and selling illegal items, such as drugs, firearms and fraudulently obtained gift cards. 

Another type of market trades in sensitive data, such as stolen credit cards, bank account details and other information that can be used for criminal activity. These sites are known as “data stores.”  

In recent years, a new breed of cyber criminals has emerged. Known as “initial access brokers,” these criminals specialize in selling access to compromised computer networks. Among their customers: ransomware gangs.  

The takedown on Tuesday of Genesis Market, a 5-year-old criminal marketplace described by officials as an “initial access broker,” offers a window into this type of cyber-criminal activity. 

It also shows how the FBI is increasingly going after users of criminal marketplaces and not just their administrators.  

U.S. officials said Genesis Market was not only a seller of stolen account access credentials but was also “one of the most prolific” initial access brokers operating on the darknet.  

Describing it as a “key enabler of ransomware,” the Justice Department said Genesis Market sold “the type of access sought by ransomware actors to attack computer networks in the United States and around the world.”  

The site went dark on Tuesday after the FBI, working with law enforcement agencies in nearly 20 countries, including the U.K. and Canada, took it offline and arrested nearly 120 people. 

In a statement, Attorney General Merrick Garland hailed the operation as “an unprecedented takedown of a major criminal marketplace that enabled cybercriminals to victimize individuals, businesses, and governments around the world.” 

Genesis is one of two popular cyber-criminal marketplaces taken down by the FBI in the past month.   

In March, the FBI shut down Breach Forums, a criminal forum and marketplace that boasted more than 340,000 members. On the Breach Forums website, users discussed tools and techniques for hacking and exploiting hacked information, according to the Justice Department. 

“We’re going after the users who leverage a service like Genesis Market, and we are doing that on a global scale,” the FBI official said. 

To take down Genesis Market, the FBI and its international law enforcement partners seized its servers and domains.  

In doing so, the FBI was able to obtain information about 59,000 individual user accounts, a senior Justice Department official said during the briefing.  

The information included usernames, passwords, email accounts, secure messenger accounts and user histories, the official said.  

“And those records helped law enforcement uncover the true identities of many of the users,” the official said.  

The users ran the gamut from online fraudsters to ransomware criminals.  

Some of the users were in the U.S., officials said, declining to provide any other details about them. They were among the 119 people arrested around the world in connection with Genesis Market takedown.  
…