Category: 3. Business

Story Continues

Experts have found weaknesses, some serious, in hundreds of tests used to check the safety and effectiveness of new artificial intelligence models being released into the world.

Computer scientists from the British government’s AI Security Institute, and experts at universities including Stanford, Berkeley and Oxford, examined more than 440 benchmarks that provide an important safety net.

They found flaws that “undermine the validity of the resulting claims”, that “almost all … have weaknesses in at least one area”, and resulting scores might be “irrelevant or even misleading”.

Many of the benchmarks are used to evaluate the latest AI models released by the big technology companies, said the study’s lead author, Andrew Bean, a researcher at the Oxford Internet Institute.

In the absence of nationwide AI regulation in the UK and US, benchmarks are used to check if new AIs are safe, align to human interests and achieve their claimed capabilities in reasoning, maths and coding.

The investigation into the tests comes amid rising concern over the safety and effectiveness of AIs, which are being released at a high pace by competing technology companies. Some have recently been forced to withdraw or tighten restrictions on AIs after they contributed to harms ranging from character defamation to suicide.

“Benchmarks underpin nearly all claims about advances in AI,” Bean said. “But without shared definitions and sound measurement, it becomes hard to know whether models are genuinely improving or just appearing to.”

Google this weekend withdrew one of its latest AIs, Gemma, after it made up unfounded allegations about a US senator having a non-consensual sexual relationship with a state trooper including fake links to news stories.

“There has never been such an accusation, there is no such individual, and there are no such new stories,” Marsha Blackburn, a Republican senator from Tennessee, told Sundar Pichai, Google’s chief executive, in a letter.

“This is not a harmless hallucination. It is an act of defamation produced and distributed by a Google-owned AI model. A publicly accessible tool that invents false criminal allegations about a sitting US senator represents a catastrophic failure of oversight and ethical responsibility.”

Google said its Gemma models were built for AI developers and researchers, not for factual assistance or for consumers. It withdrew them from its AI Studio platform after what it described as “reports of non-developers trying to use them”.

“Hallucinations – where models simply make things up about all types of things – and sycophancy – where models tell users what they want to hear – are challenges across the AI industry, particularly smaller open models like Gemma,” it said. “We remain committed to minimising hallucinations and continually improving all our models.”

Last week, Character.ai, the popular chatbot startup, banned teenagers from engaging in open-ended conversations with its AI chatbots. It followed a series of controversies, including a 14-year-old killing himself in Florida after becoming obsessed with an AI-powered chatbot that his mother claimed had manipulated him into taking his own life, and a US lawsuit from the family of a teenager who claimed a chatbot manipulated him to self-harm and encouraged him to murder his parents.

The research examined widely available benchmarks but leading AI companies also have their own internal benchmarks that were not examined.

It concluded there was a “pressing need for shared standards and best practices”.

Bean said a “shocking” finding was that only a small minority (16%) of the benchmarks used uncertainty estimates or statistical tests to show how likely a benchmark was to be accurate. In other cases where benchmarks set out to evaluate an AI’s characteristics – for example its “harmlessness” – the definition of the concept being examined was contested or ill-defined, rendering the benchmark less useful.

By Frances Yue

Still, the long-term momentum points to the upside, one technical analyst says

Bitcoin fell below its 200-day moving average on Monday.

Bitcoin has broken below an important technical level, which could signal further downside for the cryptocurrency, according to Katie Stockton, founder and managing partner at Fairlead Strategies.

Bitcoin (BTCUSD) has fallen below its 200-day moving average at $109,800, Stockton wrote in a Monday note. The 200-day moving average is one of the most widely followed indicators that may be used to define a long-term trend, and also acts as a support level for bitcoin in this case.

“We assume the corrective phase will keep hold of bitcoin for another few weeks,” based on technical indicators, Stockton wrote. Bitcoin’s next support level stands at around $94,200, she said.

Still, the long-term momentum of bitcoin remains positive, she noted, eyeing a potential price target at $134,500 in the long term if the technical trend is complete.

Bitcoin fell 3.9% on Monday to trade at around $106,400, as some large bitcoin holders sold part of their holdings, based on blockchain data.

It is unclear what triggered the move lower, analysts at crypto trading firm QCP said. “Recent selloffs, including today’s, came with no clear macro catalyst,” they wrote.

-Frances Yue

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.

  (END) Dow Jones Newswires

  11-03-25 1839ET
Copyright (c) 2025 Dow Jones & Company, Inc.

Community and environmental advocates recently won two legal challenges against fossil fuel polluters in South Africa. These resounding victories show that despite setbacks in places like the United States, international leadership and common sense can still prevail in the global fight to curb climate change.  

Earthjustice partnered with Natural Justice and the Green Connection on behalf of coastal communities fighting a plan from Total Energy and Shell to drill multiple oil and gas exploration wells off the West Coast of South Africa, roughly between Cape Town and Cape Agulhas. This region’s unique, ecologically rich, and diverse marine and coastal ecosystems not only provide habitat for numerous species but also support the cultural traditions and livelihoods of the many communities living along the coast. These communities are deeply concerned about the risks of oil spills from the exploration wells, which could reach the shoreline in as little as half a day, depending on weather conditions and the season. 

In a landmark decision in August, the Western Cape High Court set aside the project’s environmental authorization, effectively halting the project and protecting 10,000 square-kilometers (3,861 square-mi) of open ocean from the risk of catastrophic oil spills and habitat destruction that would devastate marine species.  This case has had an immediate impact because the Minister paused consideration of all other oil and gas projects until the government’s appeal of this case is decided, which could take up to six to eight months.  And if this case is upheld on appeal, it is likely that it will send all the pending oil and gas applications back to the drawing board. 

The court’s decision was a powerful rebuke to the companies and the government. First, it found that the project proposal failed to assess the cultural and economic impacts of a catastrophic oil spill on communities, which violated South Africa’s environmental review law and regulations. The court also ruled that the company’s proposal failed to consider transboundary impacts on Namibia from an oil spill.  

And importantly, the Court held that only considering the climate impacts of exploratory drilling, without considering the impacts of full-scale production drilling, was unlawful. The court agreed with our partners that exploration and production are inherently intertwined, noting, “[t]here would be no point in conducting an exploration activity unless an entity hoped to proceed to the next phase of production.”  

Another decision took the government to task for failing to adequately account for the harms caused by fossil fuels: the Supreme Court of Appeal of South Africa, the country’s second highest court, handed our partners Natural Justice, South Durban Community Environmental Alliance, and groundWork a critical victory rejecting a proposal to build a massive 3000 MW gas power plant in Richards Bay. In South Africa’s first ever legal challenge against a gas power plant, the court set aside the power plant’s environmental permit. This requires Eskom, South Africa’s state-run utility monopoly, to start from scratch if it wants to advance this ill-conceived project.  

The court found that the government and Eskom should have considered renewable energy alternatives to the proposed power plant and that “the preferred alternative must be the final choice after comparing the social, environmental, technical, and economic impacts of all the considered options” with sustainable development remaining the goal. The court also found that the government and Eskom failed to assess the cumulative impacts of extraction and transportation of the gas supplying the power plant, and that the public was denied adequate opportunities to participate in the environmental authorization process.  

This case could not have come at a more opportune time, as there are more than a dozen gas plants proposed across South Africa, and we plan to support our partners to hold the government and Eskom accountable the impacts of this gas build-out. 

Earthjustice proudly supported our partners in developing their legal arguments and look forward to continuing our work together to guarantee the right to a healthy and safe environment in South Africa, including the right to safe, clean, and affordable energy.

By Emily Bary

Third Point looked overseas to find a cheaper way to play the memory-chip boom

Third Point cheered SK Hynix’s market-leading status.

Daniel Loeb’s Third Point LLC made a bet on SK Hynix in the third quarter as it saw a way to play the artificial-intelligence trade in a cheaper way than U.S. options afforded.

His third-quarter investor letter mentioned new investments in SK Hynix (KR:000660), a South Korean memory player, and Ebara (JP:6361), a Japanese maker of semiconductor-production equipment. “We believe both companies are leaders in their respective industries, significant beneficiaries of the AI buildout, and trade at undemanding absolute valuations and meaningful discounts to their U.S. peers,” he wrote.

See also: 10 stocks that let you invest like Nvidia in the next hot AI trade

Loeb continues to see opportunities for SK Hynix, which has seen its stock surge more than 250% so far this year. The company competes against Micron Technology (MU) and Samsung Electronics (KR:005930) in the oligopolistic memory market that’s “in the early innings of de-commoditization.”

“AI workloads have been driving substantial growth in high bandwidth memory…where Hynix is the market leader with over 50% market share,” he wrote. And high-bandwidth memory confers various benefits, such as “design stickiness,” that Loeb thinks can reduce earnings volatility in the traditionally cyclical memory industry.

See also: The surprising stocks leading the tech sector this year thanks to an AI renaissance

He noted that SK Hynix is trading “at a mere 7x” estimated 2026 earnings, whereas Micron commands a 10x multiple and Samsung has a 12x multiple. Loeb also invested in SK Square (KR:402340), a “related holding company,” which has a 20% stake in SK Hynix and thereby lets investors “buy SK Hynix at a nearly 60% discount.”

Loeb also discussed the investment in Ebara, which makes chip-production equipment. The company is “a major beneficiary of the advanced packaging structures critical in AI semiconductors across both logic and memory,” he wrote, but he noted that shares trade at a “substantial discount” to rivals.

That discount reflects “an overextended cost structure as well as unnecessarily discounted pricing,” which Loeb said has hurt margins, but Third Point is “in active dialogue with Ebara’s new management team.”

U.S. artificial-intelligence investments served Third Point well in the third quarter, according to the letter, as Taiwan Semiconductor Manufacturing (TSM) (TW:2330) and Nvidia (NVDA) were its two biggest winners.

Loeb commented that OpenAI is expanding its computing capacity at a scale “difficult to comprehend” – but one nonetheless emphasizing “a compute-constrained world…that has benefited our existing investments in TSMC and Nvidia as two integral pieces in this buildout.”

Third Point’s flagship Offshore Fund returned 3.2% in the third quarter, trailing the S&P 500’s SPX 8.1% return and the MSCI World Index’s XX:990100 7.4% return.

“Our performance for the quarter and year has been below our expectations due to the weak performance of several of our largest event-driven positions including Kenvue,” the letter noted.

That stock stands to be less of a detractor in the fourth quarter, as Kimberly-Clark (KMB) announced Monday plans to purchase Kenvue (KVUE) at a 46% premium to the stock’s Friday close.

See: Could Huggies and Tylenol be a good fit? Kimberly-Clark is buying Kenvue for $48 billion

-Emily Bary

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.

  (END) Dow Jones Newswires

  11-03-25 1829ET
Copyright (c) 2025 Dow Jones & Company, Inc.

The lawsuit, filed on Friday, alleges the lithium-ion battery pack in the Model S caused the electronic door systems to fail.

Electric Vehicle company Tesla has been sued over a fiery crash in the United States that killed all five occupants of a Model S, who were allegedly trapped inside because of a design flaw that prevented them from opening the sedan’s doors.

Jeffrey Bauer, 54, and Michelle Bauer, 55, of Crandon, Wisconsin, were passengers in a Model S when the car went off the road and struck a tree in Verona, Wisconsin, a suburb of Madison, on November 1, 2024. They died the next day.

According to a complaint filed on Friday by four of the Bauers’ children, the couple’s fate was sealed because the Model S’s lithium-ion battery pack caused the electronic door systems to fail.

The children said that Tesla knew this could happen based on earlier fires, yet made a “conscious departure from known, feasible safety practices”.

Tesla, based in Austin, Texas, and led by Elon Musk, did not immediately respond to requests for comment on Monday by the Reuters news agency.

The automaker has also been sued by families of two college students killed in a Cybertruck crash last November in a San Francisco suburb, after allegedly being locked in the burning vehicle because of its door handle design.

In September, the National Highway Traffic Safety Administration disclosed a probe into the possible defects on some Tesla doors, following reports that handles could fail.

The Bauer children said that Model S rear seat passengers, like Michelle Bauer, were particularly vulnerable in the event of a crash, because they would have to lift carpeting to find a metal tab allowing their escape, which is not intuitive.

A nearby homeowner told 911 that she heard screaming from within the Bauers’ vehicle, the complaint said.

“Tesla’s design choices created a highly foreseeable risk: that occupants who survived a crash would remain trapped inside a burning vehicle,” according to the complaint.

Other defendants include the estate of the car’s driver, whom the Bauer children accused of negligent driving.

On Wall Street, Tesla’s stock finished the day up 2.5 percent.

How much power is enough for AI? Nobody knows, not even OpenAI CEO Sam Altman or Microsoft CEO Satya Nadella.

That has put software-first businesses like OpenAI and Microsoft in a bind. Much of the tech world has been focused on compute as a major barrier to AI deployment. And while tech companies have been racing to secure power, those efforts have lagged GPU purchases to the point where Microsoft has apparently ordered too many chips for the amount of power it has contracted.

“The cycles of demand and supply in this particular case you can’t really predict,” Nadella said on the BG2 podcast. “The biggest issue we are now having is not a compute glut, but it’s a power and it’s sort of the ability to get the [data center] builds done fast enough close to power.”

“If you can’t do that, you may actually have a bunch of chips sitting in inventory that I can’t plug in. In fact, that is my problem today. It’s not a supply issue of chips, it’s the fact that I don’t have warm shells to plug into,” Nadella added, referring to the commercial real estate term for buildings ready for tenants.

In some ways, we’re seeing what happens when companies accustomed to dealing with silicon and code, two technologies that scale and deploy quickly compared with massive power plants, need to ramp up their efforts in the energy world.

For more than a decade, electricity demand in the U.S. was flat. But over the last five years, demand from data centers has begun to ramp up, outpacing utilities’ plans for new generating capacity. That has led data center developers to add power in so-called behind-the-meter arrangements, where electricity is fed directly to the data center, skipping the grid.

Altman, who was also on the podcast, thinks that trouble could be brewing: “If a very cheap form of energy comes online soon at mass scale, then a lot of people are going to be extremely burned with existing contracts they’ve signed.”

“If we can continue this unbelievable reduction in cost per unit of intelligence — let’s say it’s been averaging like 40x for a given level per year — you know, that’s like a very scary exponent from an infrastructure buildout standpoint,” he said.

Altman has invested in nuclear energy, including fission startup Oklo and fusion startup Helion, along with Exowatt, a solar startup that concentrates the Sun’s heat and stores it for later use.

None of those are ready for widespread deployment today, though, and fossil-based technologies like natural gas power plants take years to build. Plus, orders placed today for new gas turbine likely won’t get fulfilled until later this decade.

That’s partially why tech companies have been adding solar at a rapid clip, drawn to the technology’s inexpensive cost, emissions-free power, and ability to deploy rapidly.

There might be subconscious factors at play, too. Photovoltaic solar is in many ways a parallel technology to semiconductors, and one that has been derisked and commoditized. Both PV solar and semiconductors are built on silicon substrates, and both roll off production lines as modular components that can be packaged together and tied into parallel arrays that make the completed part more powerful than any individual module.

Because of solar’s modularity and speed of deployment, the pace of construction is much closer to that of a data center.

But both still take time to build, and demand can change much more quickly than either a data center or solar project can be completed. Altman admitted that if AI gets more efficient or if demand doesn’t grow as he expects, some companies might be saddled with idled power plants.

But from his other comments, he doesn’t seem to think that’s likely. Instead, he appears to be a firm believer in Jevons Paradox, which says that more efficient use of a resource will lead to greater use, increasing overall demand.

“If the price of compute per like unit of intelligence or whatever — however you want to think about it — fell by a factor of a 100 tomorrow, you would see usage go up by much more than 100 and there’d be a lot of things that people would love to do with that compute that just make no economic sense at the current cost,” Altman said.

Story Continues

Operations: The company focuses on providing information infrastructure and application services in Japan. It operates with a market capitalization of ¥87.77 billion, engaging in diverse revenue streams within the tech industry.

TechMatrix, amidst a dynamic tech landscape in Asia, showcases robust growth with its earnings forecast to surge by 16.1% annually, outpacing the Japanese market’s average of 7.7%. This performance is underpinned by significant R&D investments that have catalyzed innovations across its operations. With a projected revenue increase of 11.6% per year—double the national rate—the firm leverages these advancements to enhance its competitive edge significantly. Notably, TechMatrix’s strategic focus on high-quality earnings and a strong return on equity anticipated at 20.9%, positions it well for sustained financial health and industry leadership, despite recent executive changes impacting its board structure.

Simply Wall St Growth Rating: ★★★★★☆

Overview: COVER Corporation operates a VTubers distribution platform in Japan, focusing on video and music content, with a market cap of ¥120.14 billion.

Operations: The company generates revenue primarily through its VTubers distribution platform, which focuses on video and music content in Japan. With a market capitalization of ¥120.14 billion, the business leverages digital media to engage audiences and monetize content effectively.

COVER, thriving in the high-growth tech sector in Asia, demonstrates a robust trajectory with its earnings and revenue significantly outpacing regional averages. With an annualized revenue growth rate at 14%, and earnings expanding by 20.9% per year, the company’s aggressive investment in R&D has proven fruitful, amounting to substantial yearly expenditures that fuel innovation and market competitiveness. Particularly noteworthy is COVER’s strategic pivot towards software as a service (SaaS) models, which not only enhances customer retention through subscriptions but also promises sustained revenue streams. This forward-thinking approach, coupled with a strong return on equity forecasted at 29.5%, positions COVER favorably within its industry landscape despite fierce competition and varying market dynamics.

This article by Simply Wall St is general in nature. We provide commentary based on historical data and analyst forecasts only using an unbiased methodology and our articles are not intended to be financial advice. It does not constitute a recommendation to buy or sell any stock, and does not take account of your objectives, or your financial situation. We aim to bring you long-term focused analysis driven by fundamental data. Note that our analysis may not factor in the latest price-sensitive company announcements or qualitative material. Simply Wall St has no position in any stocks mentioned.

Companies discussed in this article include KOSE:A005930 TSE:3762 and TSE:5253.

Have feedback on this article? Concerned about the content? Get in touch with us directly. Alternatively, email editorial-team@simplywallst.com