Month: April 2024

WASHINGTON — U.S. President Joe Biden was in Syracuse, New York, Thursday to tout a deal to provide memory chip maker Micron Technology with $6.1 billion in federal grants to support the firm in building factories in the states of New York and Idaho.

“We’re bringing advanced chip manufacturing back to America after 40 years,” Biden said Thursday. He said the funding, paired with a $125 billion investment from Micron, represents the “single biggest private investment ever in history of these two states.”

The investment will support the construction of two plants in Clay, a suburb of Syracuse, New York, and one in Boise, Idaho. The grant will unleash “$50 billion in private investment by 2030 as the first step towards Micron’s investment of up to $125 billion across both states over the next two decades,” the White House said in a statement.

The deal was announced last week by Senate Majority Leader Chuck Schumer, a Democrat from New York, who personally lobbied Micron to invest in his state. It’s the latest in a series of awards given by the administration, intended to shore up domestic production of advanced semiconductors using funds from the CHIPS and Science Act of 2022. The aim is to boost domestic manufacturing and reduce reliance on chip supplies from China and Taiwan.

This investment will “supercharge Micron to build the most advanced memory chip factory in the world, Schumer said Thursday. “America’s future will be built in Syracuse, not in Shanghai.”

The administration recently awarded Samsung, Taiwan Semiconductor, Intel, GlobalFoundries, Microchip Technology, and BAE Systems, more than $29 billion in federal grants for chipmaking investments. It’s part of an effort to catch up in the global semiconductor manufacturing race currently dominated by China, Taiwan and South Korea.

The U.S. share of global semiconductor manufacturing capacity has decreased from 37% in 1990 to 12% today, largely because other governments have offered manufacturing incentives and invested in research to strengthen domestic chipmaking capabilities, according to the Semiconductor Industry Association.

To address such stiff foreign competition, the $280 billion bipartisan CHIPS and Science Act offers $52 billion in incentives for domestic semiconductor production and research, as well as an investment tax credit for semiconductor manufacturing.

Manufacturing revival

The announcements are part of the economic vision the president is offering to voters in his re-election bid – that he is working to create a manufacturing revival in the country, including in Republican-controlled districts such as where the Micron plant will be located.

“Micron’s total investment will be the largest private investment in New York and Idaho’s history, and will create over 70,000 jobs, including 20,000 direct construction and manufacturing jobs and tens of thousands of indirect jobs,” the White House said.

Ahead of the November presidential election, Biden’s strategy appears to be to announce investments in manufacturing facilities in Georgia, Idaho, North Carolina and Ohio, states where Democrats lack a strong foothold.

It is not clear whether the approach will succeed as voters will not immediately feel the effects. The initial phase of the Micron project, for example, would see the first plant opened in 2028 and the second in 2029.

Meanwhile, voters are concerned about high inflation, and dislike Biden’s economic job performance. A recent Reuters/Ipsos poll shows 34% of respondents approving of Biden’s approach on the economy, compared to 41% who favor the approach of former president Donald Trump, the presumptive Republican nominee.

Still, Biden’s trip to New York is an opportunity for him to celebrate another victory following a string of good news for the president. On Wednesday, he secured the endorsement of the North America’s Building Trades Unions and signed a $95.3 billion aid package for Ukraine, Israel and Taiwan after months of congressional gridlock.

Paris Huang contributed to this report.
…

washington — The U.S. Federal Communications Commission voted 3-2 on Thursday to reinstate landmark net neutrality rules and reassume regulatory oversight of broadband internet rescinded under former President Donald Trump. 

The commission voted along party lines to finalize a proposal first advanced in October to reinstate open internet rules adopted in 2015 and re-establish the commission’s broadband authority. 

FCC Chairwoman Jessica Rosenworcel said the agency “believes every consumer deserves internet access that is fast, open, and fair.” 

“The last FCC threw this authority away and decided broadband needed no supervision,” she said. 

Net neutrality refers to the principle that internet service providers should enable access to all content and applications regardless of the source, and without favoring or blocking particular products or websites. 

The FCC said it was also using its authority to order the U.S. units of China Telecom, China Unicom and China Mobile to discontinue broadband internet access services in the United States.  

Rosenworcel noted the FCC has taken similar actions against Chinese telecom companies in the past using existing authority. 

Reinstating the net neutrality rules has been a priority for President Joe Biden, who signed a July 2021 executive order encouraging the FCC to reinstate net neutrality rules adopted under Democratic President Barack Obama. 

Democrats were stymied for nearly three years because they did not take majority control of the five-member FCC until October. 

Under Trump, the FCC had argued the net neutrality rules were unnecessary, blocked innovation and resulted in a decline in network investment by internet service providers, a contention disputed by Democrats. 

The U.S. Chamber of Commerce criticized the FCC action saying it was “imposing a flawed, pre-television era regulatory structure on broadband” and “will only deter the investments and innovation necessary to connect all Americans.” 

Public interest group Free Press said the vote is a “major victory for the public interest” saying it “empowers the FCC to hold companies like AT&T, Comcast, Spectrum and Verizon accountable for a wide range of harms to internet users across the United States.” 

A group of Republican lawmakers, including House Energy and Commerce Committee Chair Cathy McMorris Rodgers and Senator Ted Cruz, called the plan “an illegal power grab that would expose the broadband industry to an oppressive regulatory regime” giving the agency and states power to impose rate regulation, unbundle obligations and tax broadband internet providers. 

Democrats on the FCC say they will not set rate regulations. 

The Computer & Communications Industry Association, whose members include Amazon.com, Apple, Alphabet and Meta Platforms, back net neutrality, arguing the rules “must be reinstated to preserve open access to the internet.” 

USTelecom, whose members include AT&T, Verizon and others, called reinstating net neutrality “entirely counterproductive, unnecessary, and an anti-consumer regulatory distraction.” 

Despite the 2017 decision to withdraw the requirement at the federal level, a dozen states now have net neutrality laws or regulations in place. Industry groups abandoned legal challenges to those state requirements in May 2022. 
…

TOKYO — Japan’s first moon lander has survived a third freezing lunar night, Japan’s space agency said Wednesday after receiving an image from the device three months after it landed on the moon.

The Japan Aerospace Exploration Agency said the lunar probe responded to a signal from the earth Tuesday night, confirming it has survived another weekslong lunar night.

Temperatures can fall to minus 170 degrees Celsius during a lunar night and rise to around 100 Celsius during a lunar day. 

The probe, Smart Lander for Investing Moon, or SLIM, reached the lunar surface on Jan. 20, making Japan the fifth country to successfully place a probe on the moon. 

SLIM landed the wrong way up with its solar panels initially unable to see the sun, and had to be turned off within hours, but powered on when the sun rose eight days later.

SLIM, which was tasked with testing Japan’s pinpoint landing technology and collecting geological data and images, was not designed to survive lunar nights.

JAXA said on the social media platform X that SLIM’s key functions are still working despite repeated harsh cycles of temperature changes. The agency said it plans to closely monitor the lander’s deterioration. 

Scientists are hoping to find clues about the origin of the moon by comparing the mineral compositions of moon rocks and those of Earth.

The message from SLIM came days after NASA restored contact with Voyager 1, the farthest space probe from Earth, which had been sending garbled data back for months.

An U.S. lunar probe developed by a private space company announced termination of its operation a month after its February landing, while an Indian moon lander failed to establish communication after touchdown in 2023. 

 
…

Artificial intelligence surrounds U.S. political life, from fundraising to campaign advertising. Some lawmakers are looking to better police the use of generative content in this year’s presidential election as they say it threatens voter confidence in what is real. VOA correspondent Scott Stearns reports.
…

The Dutch company Holoconnects are experts in the field of holographic illusions and are now delivering life-size personal connections with a 2-meter-tall box that make it feel like the person you are talking to is physically present. Deana Mitchell has more from Austin, Texas in this week’s episode of LogOn.
…

Taiwan is looking to Southeast Asia as a pipeline to fill its shortage of high-tech talent. The numbers of foreign students coming to the island has been growing, especially from Vietnam and Indonesia. VOA Mandarin’s Peh Hong Lim reports from Hsinchu, Taiwan. Adrianna Zhang contributed.
…

LONDON — The European Union on Monday demanded TikTok provide more information about a new app that pays users to watch videos and warned that it could order the video sharing platform to suspend addictive features that pose a risk to kids. 

The 27-nation EU’s executive commission said it was opening formal proceedings to determine whether TikTok Lite breached the bloc’s new digital rules when the app was rolled out in France and Spain. 

Brussels was ratcheting up the pressure on TikTok after the company failed to respond to a request last week for information on whether the new app complies with the Digital Services Act, a sweeping law that took effect last year intending to clean up social media platforms. 

TikTok Lite is a slimmed-down version of the main TikTok app that lets users earn rewards. Points earned by watching videos, liking content and following content creators can then be exchanged for rewards including Amazon vouchers and gift cards on PayPal. 

The commission wants to see the risk assessment that TikTok should have carried out before deploying the app in the European Union. It’s worried TikTok launched the app without assessing how to mitigate “potential systemic risks” such as addictive design features that could pose harm to children. 

TikTok didn’t respond immediately to a request for comment. The company said last week it would respond to the commission’s request and noted that rewards are restricted to users 18 years and older, who have to verify their age. 

“With an endless stream of short and fast-paced videos, TikTok offers fun and a sense of connection beyond your immediate circle,” said European Commissioner Thierry Breton, one of the officials leading the bloc’s push to rein in big tech companies. “But it also comes with considerable risks, especially for our children: addiction, anxiety, depression, eating disorders, low attention spans.” 

The EU is giving TikTok 24 hours to turn over the risk assessment and until Wednesday to argue its case. Any order to suspend the TikTok Lite app’s reward features could come as early as Thursday. 

It’s the first time that the EU has issued a legally binding order for such information since the Digital Services Act took effect. Officials stepped up the pressure after TikTok failed to respond to last week’s request for the information. 

If TikTok still fails to respond, the commission warned the company also faces fines worth up to 1% of the company’s total annual income or worldwide turnover and “periodic penalties” of up to 5% of daily income or global turnover. 

TikTok was already facing intensified scrutiny from the EU. The commission already has an ongoing in-depth investigation into the main TikTok app’s DSA compliance, examining whether it’s doing enough to curb “systemic risks” stemming from its design, including “algorithmic systems” that might stimulate “behavioral addictions.” Offices are worried that measures including age verification tools to stop minors from finding “inappropriate content” might not be effective.
…

Nairobi, Kenya — Government representatives from Africa, along with ICT (information and communication technology) officials, and international organizations have gathered in Nairobi for a Connected Africa Summit. They are discussing the future of technology, unlocking the continent’s growth beyond connectivity, and addressing the challenges and opportunities in the continent’s information and technology sector.

Speaking at the Connected Africa Summit opening in Nairobi Monday, Kenyan President William Ruto said bridging the technology gap is important for Africa’s economic growth and innovation.  

“Closing the digital divide is a priority in terms of enhancing connectivity, expanding the contribution of the ICT sector to Africa’s GDP and driving overall GDP growth across all sectors. Africa’s digital economy has immense potential…,” Ruto said. “Our youth population, the youngest globally, is motivated and prepared to drive the digital economy, foster innovation and entrench new technologies.”    

Experts say digital transformation in Africa can improve its industrialization, reduce poverty, create jobs, and improve its citizens’ lives.

According to the World Bank, 36 percent of Africa’s 1.3 billion population have access to the internet, and in some of the areas that have connections, the quality of the service is poor compared to other regions.

The international financial institution figures show that Africa saw a 115 percent increase in internet users between 2016 and 2021 and that 160 million gained broadband internet access between 2019 and 2022.  

Africa’s digital growth has been hampered by the lack of an accessible, secure, and reliable internet, which is critical in closing the digital gap and reducing inequalities.  

Lacina Kone is the head of Smart Africa, an organization that coordinates ICT activities within the continent. He says integrating technology into African societies’ daily activities is necessary and cannot be ignored.  

“Digital transformation is no longer a choice but a necessity, just like water utility, just like any other utility we use at home,” Kone said. “So, this connected Africa is an opportunity for all of us. I see a lot of country members, and ICT ministers are here to align our visions together.”

The COVID-19 pandemic has accelerated the consumption of technology in different sectors of the African economy, and experts say opportunities now exist in mobile services, the development of broadband infrastructure, and data storage.  

The U.S. ambassador to Kenya, Meg Whitman, called on the summit attendees to develop technologies that can solve people’s problems.  

“I encourage all of you to consider this approach for your economies. Look at what strengths already exist in your countries and ask how technology can solve challenges in those sectors to make you a leader through innovation,” Whitman said. “Sometimes innovation looks like Artificial Intelligence, satellites and e-money. Sometimes though it looks much different than we expect. However, innovation always includes three elements: solution focused, it’s specific and it’s sustainable. Bringing solution-focused, being solution-focused is the foundation of shaping the future of a connected Africa.”

The summit ends on Friday, but before that, those attending aim to explore ways to improve Africa’s technology usage, enhance continental connectivity, boost competitiveness, and ensure the continent keeps up with the ever-evolving tech sector.
…

vancouver, british columbia — A new, digestible mini-robotic camera, about the size of a multivitamin pill, was demonstrated at the annual TED Conference in Vancouver. The remote-controlled device can eliminate invasive medical procedures.

With current technology, exploration of the digestive tract involves going through the highly invasive procedure of an endoscopy, in which a camera at the end of a cord is inserted down the throat and into a medicated patient’s stomach.

But the robotic pill, developed by Endiatx in Hayward, California, is designed to be the first motorized replacement of the procedure. A patient fasts for a day, then swallows the PillBot with lots of water. The PillBot, acting like a miniature submarine, is piloted in the body by a wireless remote control. After the exam, it then flushes out of the human body naturally.

For Dr. Vivek Kumbhari, co-founder of the company and professor of medicine and chairman of gastroenterology and hepatology at the Mayo Clinic, it is the latest step toward his goal of democratizing previously complex medicine.

If procedure-based diagnostics can be moved from a hospital to a home, “then I think we have achieved that goal,” he said. The new setting would require fewer medical staff personnel and no anesthesia, producing “a safer, more comfortable approach.”

Kumbhari said this technology also makes medicine more efficient, allowing people to get care earlier in the course of an illness.

For co-founder Alex Luebke, the micro-robotic pill can be transformative for rural areas around the world where there is limited access to medical facilities.

“Especially in developing countries, there is no access” to complex medical procedures, he said. “So being able to have the technology, gather all that information and provide you the solution, even in remote areas – that’s the way to do it.”

Luebke said if internet access is not immediately available, information from the PillBot can be transmitted later.

The duo are also utilizing artificial intelligence to provide the initial diagnosis, with a medical doctor later developing a treatment plan.

Joel Bervell is known to his million social media followers as the “Medical Mythbuster” and is a fourth-year medical student at Washington State University. He said the strength of this type of technology is how it can be easily used in remote and rural communities.

Many patients “travel hundreds of miles, literally, for their appointment. Use of a pill that would not require a visit to a physician “would be life-changing for them.” 

The micro-robotic pill is undergoing trials and will soon be in front of the U.S. Food and Drug Administration for approval, which developers expect to have in 2025. It’s expected that the pill would then be widely available in 2026.

Kumbhari hopes the technology can be expanded to the bowels, vascular system, heart, liver, brain and other parts of the body. Eventually, he hopes, this will allow hospitals to be left for more urgent medical care and surgeries.
…

HONG KONG — Apple said it had removed Meta’s WhatsApp messaging app and its Threads social media app from the App Store in China to comply with orders from Chinese authorities.

The apps were removed from the store Friday after Chinese officials cited unspecified national security concerns.

Their removal comes amid elevated tensions between the U.S. and China over trade, technology and national security.

The U.S. has threatened to ban TikTok over national security concerns. But while TikTok, owned by Chinese technology firm ByteDance, is used by millions in the U.S., apps like WhatsApp and Threads are not commonly used in China.

Instead, the messaging app WeChat, owned by Chinese company Tencent, reigns supreme.

Other Meta apps, including Facebook, Instagram and Messenger remained available for download, although use of such foreign apps is blocked in China due to its “Great Firewall” network of filters that restrict use of foreign websites such as Google and Facebook.

“The Cyberspace Administration of China ordered the removal of these apps from the China storefront based on their national security concerns,” Apple said in a statement.

“We are obligated to follow the laws in the countries where we operate, even when we disagree,” Apple said.

A spokesperson for Meta referred to “Apple for comment.”

Apple, previously the world’s top smartphone maker, recently lost the top spot to Korean rival Samsung Electronics. The U.S. firm has run into headwinds in China, one of its top three markets, with sales slumping after Chinese government agencies and employees of state-owned companies were ordered not to bring Apple devices to work.

Apple has been diversifying its manufacturing bases outside China.

Its CEO Tim Cook has been visiting Southeast Asia this week, traveling to Hanoi and Jakarta before wrapping up his travels in Singapore. On Friday he met with Singapore’s deputy prime minister, Lawrence Wong, where they “discussed the partnership between Singapore and Apple, and Apple’s continued commitment to doing business in Singapore.”

Apple pledged to invest over $250 million to expand its campus in the city-state.

Earlier this week, Cook met with Vietnamese Prime Minister Pham Minh Chinh in Hanoi, pledging to increase spending on Vietnamese suppliers.

He also met with Indonesian President Joko Widodo. Cook later told reporters that they talked about Widodo’s desire to promote manufacturing in Indonesia, and said that this was something that Apple would “look at.”
…

Sundsvall,  Sweden — German Chancellor Olaf Scholz’s short videos of his three-day trip to China this week proved popular in posts on Chinese-owned social media platform TikTok, which the European Union, Canada, Taiwan and the United States banned on official devices more than a year ago, citing security concerns.

By Friday, one video showing highlights of Scholz’s trip had garnered 1.5 million views while another of him speaking about it on the plane home had 1.4 million views. 

Scholz opened his TikTok account April 8 to attract youth, promising he wouldn’t post videos of himself dancing.  His most popular post so far, about his 40-year-old briefcase, was watched 3.6 million times.  Many commented, “This briefcase is older than me.”

Scholtz is one of several Western leaders to use TikTok, despite concerns that its parent company, ByteDance, could provide private user data to the Chinese government and could also be used to push a pro-Beijing agenda.

Greek Prime Minister Kyriakos Mitsotakis has 258,000 followers on TikTok, and Irish Prime Minister Simon Harris has 99,000 followers. 

U.S. President Joe Biden’s reelection campaign team opened a TikTok account in February, despite Biden himself vowing to sign legislation expected to be voted on as early as Saturday to force ByteDance to divest in the U.S. or face a ban. 

Former U.S. President Donald Trump, who unsuccessfully tried to ban TikTok in 2020, in March reversed his position and now appears to oppose a ban. 

ByteDance denies it would provide user data to the Chinese government, despite reports indicating it could be at risk, and China has firmly opposed any forced sale.

Kevin Morgan, TikTok’s director of security and integrity in Europe, the Middle East and Africa, says more than 134 million people in 27 EU countries visit TikTok every month, including a third of EU lawmakers. 

As the European Union’s June elections approach, more European politicians are using the popular platform favored by young people to attract votes. 

Ola Patrik Bertil Moeller, a Swedish legislator with the Social Democratic Party who has 124,000 followers on TikTok, told VOA, “We as politicians participate in the conversation and spread accurate images and answer the questions that people have. If we’re not there, other forces that don’t want good will definitely be there.”

But other European politicians see TikTok as risky.  

Norwegian Prime Minister Jonas Gahr Store on Monday expressed his uneasiness about social media platforms, including TikTok, being “used by various threat actors for several purposes, such as recruitment for espionage, influencing through disinformation and fake news, or mapping regime critics. This is disturbing.”

Konstantin von Notz, vice-chairman of the Green Parliamentary Group in the German legislature, told VOA, “While questions of security and the protection of personal data generally arise when using social networks, the issue is even more relevant for users of TikTok due to the company’s proximity to the Chinese state.” 

Matthias C. Kettemann, an internet researcher at the Leibniz Institute for Media Research in Hamburg, Germany, told VOA, “Keeping data safe is a difficult task; given TikTok’s ties to China doesn’t make it easier.”  But he emphasized, “TikTok is obliged to do these measures through the EU’s GDPR [General Data Protection Regulation] anyway from a legal side.”

But analysts question whether ByteDance will obey European law if pressed by the Chinese state.

Matthias Spielkamp, executive director AlgorithmWatch, told VOA, “Does TikTok have an incentive to comply with European law? Yes, there’s an enormous amount of money on the line. Is it realistic that TikTok, being owned by a Chinese company, can resist requests for data by its Chinese parent? Hardly. How is this going to play out? No one knows right now.”

Adrianna Zhang contributed to this report.
…

CAMBRIDGE, Massachusetts — Facebook parent Meta Platforms has unveiled a new set of artificial intelligence systems that are powering what CEO Mark Zuckerberg calls “the most intelligent AI assistant that you can freely use.” 

But as Zuckerberg’s crew of amped-up Meta AI agents started venturing into social media in recent days to engage with real people, their bizarre exchanges exposed the ongoing limitations of even the best generative AI technology. 

One joined a Facebook moms group to talk about its gifted child. Another tried to give away nonexistent items to confused members of a Buy Nothing forum. 

Meta, along with leading AI developers Google and OpenAI, and startups such as Anthropic, Cohere and France’s Mistral, have been churning out new AI language models and hoping to convince customers they’ve got the smartest, handiest or most efficient chatbots. 

While Meta is saving the most powerful of its AI models, called Llama 3, for later, on Thursday it publicly released two smaller versions of the same Llama 3 system and said it’s now baked into the Meta AI assistant feature in Facebook, Instagram and WhatsApp. 

AI language models are trained on vast pools of data that help them predict the most plausible next word in a sentence, with newer versions typically smarter and more capable than their predecessors. Meta’s newest models were built with 8 billion and 70 billion parameters — a measurement of how much data the system is trained on. A bigger, roughly 400 billion-parameter model is still in training. 

“The vast majority of consumers don’t candidly know or care too much about the underlying base model, but the way they will experience it is just as a much more useful, fun and versatile AI assistant,” Nick Clegg, Meta’s president of global affairs, said in an interview. 

‘A little stiff’

He added that Meta’s AI agent is loosening up. Some people found the earlier Llama 2 model — released less than a year ago — to be “a little stiff and sanctimonious sometimes in not responding to what were often perfectly innocuous or innocent prompts and questions,” he said. 

But in letting down their guard, Meta’s AI agents have also been spotted posing as humans with made-up life experiences. An official Meta AI chatbot inserted itself into a conversation in a private Facebook group for Manhattan moms, claiming that it, too, had a child in the New York City school district. Confronted by group members, it later apologized before the comments disappeared, according to a series of screenshots shown to The Associated Press. 

“Apologies for the mistake! I’m just a large language model, I don’t have experiences or children,” the chatbot told the group. 

One group member who also happens to study AI said it was clear that the agent didn’t know how to differentiate a helpful response from one that would be seen as insensitive, disrespectful or meaningless when generated by AI rather than a human. 

“An AI assistant that is not reliably helpful and can be actively harmful puts a lot of the burden on the individuals using it,” said Aleksandra Korolova, an assistant professor of computer science at Princeton University. 

Clegg said Wednesday that he wasn’t aware of the exchange. Facebook’s online help page says the Meta AI agent will join a group conversation if invited, or if someone “asks a question in a post and no one responds within an hour.” The group’s administrators have the ability to turn it off. 

Need a camera?

In another example shown to the AP on Thursday, the agent caused confusion in a forum for swapping unwanted items near Boston. Exactly one hour after a Facebook user posted about looking for certain items, an AI agent offered a “gently used” Canon camera and an “almost-new portable air conditioning unit that I never ended up using.” 

Meta said in a written statement Thursday that “this is new technology and it may not always return the response we intend, which is the same for all generative AI systems.” The company said it is constantly working to improve the features. 

In the year after ChatGPT sparked a frenzy for AI technology that generates human-like writing, images, code and sound, the tech industry and academia introduced 149 large AI systems trained on massive datasets, more than double the year before, according to a Stanford University survey. 

They may eventually hit a limit, at least when it comes to data, said Nestor Maslej, a research manager for Stanford’s Institute for Human-Centered Artificial Intelligence. 

“I think it’s been clear that if you scale the models on more data, they can become increasingly better,” he said. “But at the same time, these systems are already trained on percentages of all the data that has ever existed on the internet.” 

More data — acquired and ingested at costs only tech giants can afford, and increasingly subject to copyright disputes and lawsuits — will continue to drive improvements. “Yet they still cannot plan well,” Maslej said. “They still hallucinate. They’re still making mistakes in reasoning.” 

Getting to AI systems that can perform higher-level cognitive tasks and common-sense reasoning — where humans still excel— might require a shift beyond building ever-bigger models. 

Seeing what works

For the flood of businesses trying to adopt generative AI, which model they choose depends on several factors, including cost. Language models, in particular, have been used to power customer service chatbots, write reports and financial insights, and summarize long documents. 

“You’re seeing companies kind of looking at fit, testing each of the different models for what they’re trying to do and finding some that are better at some areas rather than others,” said Todd Lohr, a leader in technology consulting at KPMG. 

Unlike other model developers selling their AI services to other businesses, Meta is largely designing its AI products for consumers — those using its advertising-fueled social networks. Joelle Pineau, Meta’s vice president of AI research, said at a recent London event that the company’s goal over time is to make a Llama-powered Meta AI “the most useful assistant in the world.” 

“In many ways, the models that we have today are going to be child’s play compared to the models coming in five years,” she said. 

But she said the “question on the table” is whether researchers have been able to fine-tune its bigger Llama 3 model so that it’s safe to use and doesn’t, for example, hallucinate or engage in hate speech. In contrast to leading proprietary systems from Google and OpenAI, Meta has so far advocated for a more open approach, publicly releasing key components of its AI systems for others to use. 

“It’s not just a technical question,” Pineau said. “It is a social question. What is the behavior that we want out of these models? How do we shape that? And if we keep on growing our model ever more in general and powerful without properly socializing them, we are going to have a big problem on our hands.”
…