An anonymous reader quotes a report from The Guardian: The Danish government is to clamp down on the creation and dissemination of AI-generated deepfakes by changing copyright law to ensure that everybody has the right to their own body, facial features and voice. The Danish government said on Thursday it would strengthen protection against digital imitations of people's identities with what it believes to be the first law of its kind in Europe. Having secured broad cross-party agreement, the department of culture plans to submit a proposal to amend the current law for consultation before the summer recess and then submit the amendment in the autumn. It defines a deepfake as a very realistic digital representation of a person, including their appearance and voice.

The Danish culture minister, Jakob Engel-Schmidt, said he hoped the bill before parliament would send an "unequivocal message" that everybody had the right to the way they looked and sounded. He told the Guardian: "In the bill we agree and are sending an unequivocal message that everybody has the right to their own body, their own voice and their own facial features, which is apparently not how the current law is protecting people against generative AI." He added: "Human beings can be run through the digital copy machine and be misused for all sorts of purposes and I'm not willing to accept that."

The changes to Danish copyright law will, once approved, theoretically give people in Denmark the right to demand that online platforms remove such content if it is shared without consent. It will also cover "realistic, digitally generated imitations" of an artist's performance without consent. Violation of the proposed rules could result in compensation for those affected. The government said the new rules would not affect parodies and satire, which would still be permitted. "Of course this is new ground we are breaking, and if the platforms are not complying with that, we are willing to take additional steps," said Engel-Schmidt.

He expressed hope that other European countries will follow suit and warned that "severe fines" will be imposed if tech platforms fail to comply.

36% of all Chinese undergraduate entrants -- about 1.6 million people -- selected engineering degrees in 2022 (the latest year for which data are available), up from 32% in 2010, according to data from China's Ministry of Education. In Britain and America, which have far fewer students to start with, the proportion hovers around 5%.

The surge comes as China's government directs universities to focus on strategic industries and technological bottlenecks. Over 600 Chinese universities now offer undergraduate programs in artificial intelligence, a field the Communist Party vows to dominate by 2030. In 2023, officials started telling universities to overhaul their degree programs, and the education ministry announced an "emergency mechanism" to create degrees more quickly to meet "national priorities." Over half of China's young people now complete some form of higher education through 3,000-odd institutions. Youth unemployment reached 14.9% in May, driving students toward technical fields they believe offer better job prospects.

The UK government has rejected the 25 billion ($34.39 billion) pound Morocco-UK Power Project, citing a preference for domestic renewable initiatives that offer greater economic and strategic benefits. The project aimed to supply solar and wind energy from the Sahara to power up to seven million UK homes. Reuters reports: "The government has concluded that it is not in the UK national interest at this time to continue further consideration of support for the Morocco-UK Power Project," energy department minister Michael Shanks said in a written statement to parliament. He also said the project did not clearly align strategically with the government's mission to build homegrown power in the UK.

Xlinks' Morocco-UK power project would have tapped Moroccan renewable energy via what would have been the world's longest subsea power cable. The plan involved building 3,800 kilometers (2,361 miles) of high-voltage direct current subsea cables from Morocco to southwest England. The company had been seeking a guaranteed minimum price for the electricity supplied, known as contract for difference, from Britain's government.

An anonymous reader quotes a report from the Associated Press: The head of the federal government agency that oversees Fannie Mae and Freddie Mac wants the mortgage giants to consider accepting a homebuyer's cryptocurrency holdings in their criteria for buying mortgages from banks. William Pulte, director of the Federal Housing Finance Agency, which oversees Fannie and Freddie, ordered the agencies Wednesday to prepare a proposal for consideration of crypto as an asset for reserves when they assess risks in single-family home loans.

Pulte also instructed the agencies that their mortgage risk assessments should not require cryptocurrency assets to be converted to U.S. dollars. And only crypto assets that "can be evidenced and stored on a U.S.-regulated centralized exchange subject to all applicable laws" are to be considered by the agencies in their proposal, Pulte wrote in a written order, effective immediately. Pulte was sworn in as the head of FHFA in March. Public records show that as of January 2025, Pulte's spouse owned between $500,000 and $1 million of bitcoin and a similar amount of Solana's SOL token. [...]

The policy change is meant to encourage banks to expand how they gauge borrowers' creditworthiness, in hopes that more aspiring homebuyers can qualify for a home loan. It also recognizes that cryptocurrencies have grown in popularity as an alternative to traditional investments, such as bonds and stocks. The agencies have to come up with their proposals "as soon as reasonably practical," according to the order. "This is a big win for advocates of cryptocurrencies who want crypto to be treated the same way as other assets are," said Daryl Fairweather, chief economist at Redfin.

Currently, stock investments are treated as qualifying assets that count toward reserves that banks want borrowers to have. But assets that are more volatile, like individual stocks or crypto, may be discounted by lenders, Fairweather noted. "As long as lenders are appropriately discounting crypto based on volatility, it's fine that crypto investments count toward reserves," she said.

Danielle Hale, chief economist at Realtor.com, added: "If Fannie and Freddie are going to accept cryptocurrency as collateral, that's a strong incentive for banks to shift their practices. Because people who might otherwise have to sell cryptocurrency to qualify -- and maybe that's a deal-breaker for them now -- under this new policy, they can qualify. It sort of expands the potential pool of eligible buyers."

U.S. lawmakers have reintroduced the bipartisan Open App Markets Act, aiming to curb Apple and Google's control over mobile app stores by promoting competition, supporting third-party marketplaces and sideloading, and safeguarding developer rights. AppleInsider reports: The Open App Markets Act seeks to do a number of things, including:
- Protect developers' rights to tell consumers about lower prices and offer competitive pricing;
- Protect sideloading of apps;
- Promote competition by opening the market to third-party app stores, startup apps, and alternative payment systems;
- Make it possible for developers to offer new experiences that take advantage of consumer device features;
- Give consumers greater control over their devices;
- Prevent app stores from disadvantaging developers; and
- Establish safeguards to preserve consumer privacy, security, and safety.

This isn't the first time we've seen this bill, either. In 2021, Senators Blumenthal, Klobuchar, and Blackburn had attempted to put forth the original version of the Open App Markets Act.However, the initial bill never made it to the floor for an office vote. Thanks to last-minute efforts by lobbying groups and appearances from chief executives, the bill eventually stalled out.

While the two bills are largely similar, the revised version introduces several key differences. Notably, the new version includes new carve-outs aimed at protecting intellectual property and addressing potential national security concerns.There's also a new clause that would prohibit punitive actions against developers for enabling remote access to other apps. The clause addition harkens back to the debacle between Apple and most game streaming services -- though in 2024, Apple loosened its App Store guidelines to allow cloud gaming and emulation.

There are a few new platform-protective clauses added, too. For instance, it would significantly lower the burden of proof for either Apple or Google to block platform access to a third-party app.Additionally, it reinforces the fact that companies like Apple or Google will not need to provide support or refunds for third-party apps installed outside of first-party app marketplaces. The full bill can be found here.

Australia's eSafety Commissioner has urged the government to deny YouTube an exemption from upcoming child safety regulations, citing research showing it exposes more children to harmful content than any other platform. YouTube pushed back, calling the commissioner's stance inconsistent with government data and parental feedback. "The quarrel adds an element of uncertainty to the December rollout of a law being watched by governments and tech leaders around the world as Australia seeks to become the first country to fine social media firms if they fail to block users aged under 16," reports Reuters. From the report: The centre-left Labor government of Anthony Albanese has previously said it would give YouTube a waiver, citing the platform's use for education and health. Other social media companies such as Meta's Facebook and Instagram, Snapchat, and TikTok have argued such an exemption would be unfair. eSafety Commissioner Julie Inman Grant said she wrote to the government last week to say there should be no exemptions when the law takes effect. She added that the regulator's research found 37% of children aged 10 to 15 reported seeing harmful content on YouTube -- the most of any social media site. [...]

YouTube, in a blog post, accused Inman Grant of giving inconsistent and contradictory advice, which discounted the government's own research which found 69% of parents considered the video platform suitable for people under 15. "The eSafety commissioner chose to ignore this data, the decision of the Australian Government and other clear evidence from teachers and parents that YouTube is suitable for younger users," wrote Rachel Lord, YouTube's public policy manager for Australia and New Zealand.

Inman Grant, asked about surveys supporting a YouTube exemption, said she was more concerned "about the safety of children and that's always going to surpass any concerns I have about politics or being liked or bringing the public onside". A spokesperson for Communications Minister Anika Wells said the minister was considering the online regulator's advice and her "top priority is making sure the draft rules fulfil the objective of the Act and protect children from the harms of social media."

Health and Human Services Secretary Robert F. Kennedy Jr. announced a major federal campaign to promote wearable health tech, aiming for every American to adopt a device within four years as part of a broader effort to "Make America Healthy Again." Gizmodo reports: RFK Jr. announced the initiative Tuesday afternoon during a House Energy and Commerce Health Subcommittee meeting to discuss the HHS' budget request for the upcoming fiscal year. In response to a question from representative Troy Balderson (R-Ohio) about wearables, Kennedy revealed that HHS will soon conduct one of the agency's largest ever advertising campaigns to promote their use. He added that in his ideal future, every American will be donning a wearable within the next four years. "It's a key part of our mission to Make America Healthy Again," RFK Jr. stated in an X post following the question.

China's advantages in developing AI are about to unleash a wave of innovation that will generate more than 100 DeepSeek-like breakthroughs in the coming 18 months, according to a former top official. From a report: The new software products "will fundamentally change the nature and the tech nature of the whole Chinese economy," Zhu Min, who was previously a deputy governor of the People's Bank of China, said during the World Economic Forum in Tianjin on Tuesday.

Zhu, who also served as the deputy managing director at the International Monetary Fund, sees a transformation made possible by harnessing China's pool of engineers, massive consumer base and supportive government policies. The bullish take on China's AI future promises no letup in the competition for dominance in cutting-edge technologies with the US, just as the world's two biggest economies are also locked in a trade war.

An anonymous reader quotes a report from OilPrice.com: China installed its highest solar power capacity for a single month in May, according to official data, which showed mind-boggling figures that the country installed more solar capacity in a month than any other nation did for the entire 2024. With 93 gigawatts (GW) of solar capacity installed in May, China smashed its own record of 71 GW in December 2024, per data from the National Energy Administration cited by Bloomberg.

China's solar capacity additions in May were rushed ahead of a new government policy -- effective June 1 -- to remove pricing protection for solar power projects. Under these protections, solar projects had all but guaranteed profits when they start operations. Another new rule, effective May 1, made connecting rooftop panels to the grid more difficult. These new policies are expected to moderate the growth in solar power additions this summer, analysts say. A separate report notes that China's cumulative installed solar capacity has surpassed 1 TW, according to the National Energy Administration (NEA). "By the end of May 2025, solar capacity had reached 1.08 TW (1,080 GW), up 56.9% year on year," reports pv magazine.

"NEA data show total power generation capacity stood at 3.61 TW at the end of May, an 18.8% increase from a year earlier."

Hackers suspected of working on behalf of the Chinese government exploited a maximum-severity vulnerability, which had received a patch 16 months earlier, to compromise a telecommunications provider in Canada, officials from that country and the US said Monday. ArsTechnica: "The Cyber Centre is aware of malicious cyber activities currently targeting Canadian telecommunications companies," officials for the center, the Canadian government's primary cyber security agency, said in a statement. "The responsible actors are almost certainly PRC state-sponsored actors, specifically Salt Typhoon." The FBI issued its own nearly identical statement.

Salt Typhoon is the name researchers and government officials use to track one of several discreet groups known to hack nations all over the world on behalf of the People's Republic of China. In October 2023, researchers disclosed that hackers had backdoored more than 10,000 Cisco devices by exploiting CVE-2023-20198, a vulnerability with a maximum severity rating of 10. Any switch, router, or wireless LAN controller running Cisco's iOS XE that had the HTTP or HTTPS server feature enabled and exposed to the Internet was vulnerable. Cisco released a security patch about a week after security firm VulnCheck published its report.

An anonymous reader shares a report: AI firm DeepSeek is aiding China's military and intelligence operations, a senior U.S. official told Reuters, adding that the Chinese tech startup sought to use Southeast Asian shell companies to access high-end semiconductors that cannot be shipped to China under U.S. rules. The U.S. conclusions reflect a growing conviction in Washington that the capabilities behind the rapid rise of one of China's flagship AI enterprises may have been exaggerated and relied heavily on U.S. technology.

[...] "We understand that DeepSeek has willingly provided and will likely continue to provide support to China's military and intelligence operations," a senior State Department official told Reuters in an interview. "This effort goes above and beyond open-source access to DeepSeek's AI models," the official said, speaking on condition of anonymity in order to speak about U.S. government information. Chinese law requires companies operating in China to provide data to the government when requested. But the suggestion that DeepSeek is already doing so is likely to raise privacy and other concerns for the firm's tens of millions of daily global users.

The U.S. House chief administrative officer has banned WhatsApp from congressional staffers' government devices citing data vulnerability concerns. The cybersecurity office deemed the messaging app "high-risk" due to lack of transparency in data protection, absence of stored data encryption, and potential security risks, according to an email obtained by Axios.

Staff cannot download or keep WhatsApp on any House device, including mobile, desktop, or web browser versions.

An anonymous reader shared this report from the Washington Post: Over the past three years, companies have invested tens of billions of dollars toward making electric vehicles in the United States, buoyed by tax incentives aimed at helping American businesses compete with China. Now, those companies are facing a strange problem: too much manufacturing capacity, not enough demand.

As sales of electric vehicles slow and congressional Republicans take aim at EV tax credits and incentives, the United States is slated to have more battery and EV manufacturing than it needs, according to a report released Wednesday by the Rhodium Group, a research firm. That could leave factories — many of which are already operating or under construction — stranded if car sales continue to slump. "The rug is being pulled out from under these manufacturers," said Hannah Pitt, a director in Rhodium's energy and climate practice...

After [America's 2022 climate bill], battery investment in the U.S. skyrocketed. Companies went from investing about $1 billion per quarter in 2022 to $11 billion per quarter in 2024. Most of that battery investment went to red states, including in the South's "Battery Belt," where manufacturers were drawn to inexpensive land and a nonunionized workforce. Now, however, that battery boom is teetering. In the first three months of 2025, companies canceled $6 billion in battery manufacturing — a record. EV sales have slowed...

According to the new report, the United States has almost enough battery capacity announced or under development to meet demand all the way to 2030 if EV sales continue to slump. That might sound like a good thing — but if EV sales drop further, it means companies will be left with factories they won't be able to use. At the same time, China has excess battery capacity. The country has enough manufacturing to meet the entire world's demand for batteries — and may be looking to off-load them onto other markets... And if the incentives for using U.S.-made batteries disappear, the nation's manufacturers would be left high and dry.

An anonymous reader quotes a report from Reuters: SoftBank Group founder Masayoshi Son is envisaging setting up a $1 trillion industrial complex in Arizona that will build robots and artificial intelligence, Bloomberg News reported on Friday, citing people familiar with the matter. Son is seeking to team up with Taiwan Semiconductor Manufacturing Co for the project, which is aimed at bringing back high-end tech manufacturing to the U.S. and to create a version of China's vast manufacturing hub of Shenzhen, the report said.

SoftBank officials have spoken with U.S. federal and state government officials to discuss possible tax breaks for companies building factories or otherwise investing in the industrial park, including talks with U.S. Secretary of Commerce Howard Lutnick, the report said. SoftBank is keen to have TSMC involved in the project, codenamed Project Crystal Land, but it is not clear in what capacity, the report said. It is also not clear the Taiwanese company would be interested, it said. TSMC is already building chipmaking factories in the U.S. with a planned investment of $165 billion. Son is also sounding out interest among tech companies including Samsung Electronics, the report said.

The plans are preliminary and feasibility depends on support from the Trump administration and state officials, it said. A commitment of $1 trillion would be double that of the $500 billion "Stargate" project which seeks to build out data centre capacity across the U.S., with funding from SoftBank, OpenAI and Oracle.

Ancient Slashdot reader Alain Williams shares a report from The Guardian: The BBC is threatening legal action against Perplexity AI, in the corporation's first move to protect its content from being scraped without permission to build artificial intelligence technology. The corporation has sent a letter to Aravind Srinivas, the chief executive of the San Francisco-based startup, saying it has gathered evidence that Perplexity's model was "trained using BBC content." The letter, first reported by the Financial Times, threatens an injunction against Perplexity unless it stops scraping all BBC content to train its AI models, and deletes any copies of the broadcaster's material it holds unless it provides "a proposal for financial compensation."

The legal threat comes weeks after Tim Davie, the director general of the BBC, and the boss of Sky both criticised proposals being considered by the government that could let tech companies use copyright-protected work without permission. "If we currently drift in the way we are doing now we will be in crisis," Davie said, speaking at the Enders conference. "We need to make quick decisions now around areas like ... protection of IP. We need to protect our national intellectual property, that is where the value is. What do I need? IP protection; come on, let's get on with it." "Perplexity's tool [which allows users to choose between different AI models] directly competes with the BBC's own services, circumventing the need for users to access those services," the corporation said.

Perplexity told the FT that the BBC's claims were "manipulative and opportunistic" and that it had a "fundamental misunderstanding of technology, the internet and intellectual property law."

Australia's world-first social media ban for under-16s moved closer to implementation after a key trial found that checking a user's age is technologically possible and can be integrated into existing services. From a report: The conclusions are a blow to Facebook-owner Meta Platforms, TikTok and Snap, which opposed the controversial legislation. Some platform operators had questioned whether a user's age could be reliably established using current technology.

The results of the government-backed trial clear the way for the law to come into force by the end of the year. The findings also potentially allow other jurisdictions to follow Australia's lead as countries around the world grapple with ways to protect children from harmful content online. "Age assurance can be done in Australia and can be private, robust and effective," the government-commissioned Age Assurance Technology Trial said in a statement Friday announcing its preliminary findings.

An anonymous reader quotes a report from Cybernews: Several collections of login credentials reveal one of the largest data breaches in history, totaling a humongous 16 billion exposed login credentials. The data most likely originates from various infostealers. Unnecessarily compiling sensitive information can be as damaging as actively trying to steal it. For example, the Cybernews research team discovered a plethora of supermassive datasets, housing billions upon billions of login credentials. From social media and corporate platforms to VPNs and developer portals, no stone was left unturned.

Our team has been closely monitoring the web since the beginning of the year. So far, they've discovered 30 exposed datasets containing from tens of millions to over 3.5 billion records each. In total, the researchers uncovered an unimaginable 16 billion records. None of the exposed datasets were reported previously, bar one: in late May, Wired magazine reported a security researcher discovering a "mysterious database" with 184 million records. It barely scratches the top 20 of what the team discovered. Most worryingly, researchers claim new massive datasets emerge every few weeks, signaling how prevalent infostealer malware truly is.

"This is not just a leak -- it's a blueprint for mass exploitation. With over 16 billion login records exposed, cybercriminals now have unprecedented access to personal credentials that can be used for account takeover, identity theft, and highly targeted phishing. What's especially concerning is the structure and recency of these datasets -- these aren't just old breaches being recycled. This is fresh, weaponizable intelligence at scale," researchers said. The only silver lining here is that all of the datasets were exposed only briefly: long enough for researchers to uncover them, but not long enough to find who was controlling vast amounts of data. Most of the datasets were temporarily accessible through unsecured Elasticsearch or object storage instances. Key details to be aware of: - The records include billions of login credentials, often structured as URL, login, and password.
- The datasets include both old and recent breaches, many with cookies, tokens, and metadata, making them especially dangerous for organizations without multi-factor authentication or strong credential practices.
- Exposed services span major platforms like Apple, Google, Facebook, Telegram, GitHub, and even government services.
- The largest dataset alone includes 3.5 billion records, while one associated with the Russian Federation has over 455 million; many dataset names suggest links to malware or specific regions.
- Ownership of the leaked data is unclear, but its potential for phishing, identity theft, and ransomware is severe -- especially since even a - Basic cyber hygiene -- such as regularly updating strong passwords and scanning for malware -- is currently the best line of defense for users.

An anonymous reader quotes a report from The Guardian: Foreign students will be required to unlock their social media profiles to allow US diplomats to review their online activity before receiving educational and exchange visas, the state department has announced. Those who fail to do so will be suspected of hiding that activity from US officials. The new guidance, unveiled by the state department on Wednesday, directs US diplomats to conduct an online presence review to look for "any indications of hostility toward the citizens, culture, government, institutions, or founding principles of the United States."

A cable separately obtained by Politico also instructs diplomats to flag any "advocacy for, aid or support for foreign terrorists and other threats to US national security" and "support for unlawful antisemitic harassment or violence." The screening for "antisemitic" activity matches similar guidance given at US Citizenship and Immigration Services under the Department of Homeland Security and has been criticized as an effort to crack down on opposition to the conduct of Israel's war in Gaza.

The new state department checks are directed at students and other applicants for visas in the F, M and J categories, which refer to academic and vocational education, as well as cultural exchanges. "It is an expectation from American citizens that their government will make every effort to make our country safer, and that is exactly what the Trump administration is doing every single day," said a senior state department official, adding that Marco Rubio was "helping to make America and its universities safer while bringing the state Department into the 21st century."

Austria's coalition government has agreed on a plan to enable police to monitor suspects' secure messaging in order to thwart militant attacks, ending what security officials have said is a rare and dangerous blind spot for a European Union country. From a report: Because Austria lacks a legal framework for monitoring messaging services like WhatsApp, its main domestic intelligence service and police rely on allies with far more sweeping powers like Britain and the United States alerting them to chatter about planned attacks and spying.

That kind of tip-off led to police unravelling what they say was a planned attack on a Taylor Swift concert in Vienna, which prompted the cancellation of all three of her planned shows there in August of last year. "The aim is to make people planning terrorist attacks in Austria feel less secure - and increase everyone else's sense of security," Joerg Leichtfried of the Social Democrats, the junior minister in charge of overseeing the Directorate for State Security and Intelligence (DSN), told a news conference.

An anonymous reader quotes a report from Time Magazine: While AI could offer transformative benefits, without proper safeguards it could facilitate nuclear and biological threats and cause "potentially irreversible harms," a new report commissioned by California Governor Gavin Newsom has warned. "The opportunity to establish effective AI governance frameworks may not remain open indefinitely," says the report, which was published on June 17 (PDF). Citing new evidence that AI can help users source nuclear-grade uranium and is on the cusp of letting novices create biological threats, it notes that the cost for inaction at this current moment could be "extremely high." [...]

"Foundation model capabilities have rapidly advanced since Governor Newsom vetoed SB 1047 last September," the report states. The industry has shifted from large language AI models that merely predict the next word in a stream of text toward systems trained to solve complex problems and that benefit from "inference scaling," which allows them more time to process information. These advances could accelerate scientific research, but also potentially amplify national security risks by making it easier for bad actors to conduct cyberattacks or acquire chemical and biological weapons. The report points to Anthropic's Claude 4 models, released just last month, which the company said might be capable of helping would-be terrorists create bioweapons or engineer a pandemic. Similarly, OpenAI's o3 model reportedly outperformed 94% of virologists on a key evaluation. In recent months, new evidence has emerged showing AI's ability to strategically lie, appearing aligned with its creators' goals during training but displaying other objectives once deployed, and exploit loopholes to achieve its goals, the report says. While "currently benign, these developments represent concrete empirical evidence for behaviors that could present significant challenges to measuring loss of control risks and possibly foreshadow future harm," the report says.

While Republicans have proposed a 10 year ban on all state AI regulation over concerns that a fragmented policy environment could hamper national competitiveness, the report argues that targeted regulation in California could actually "reduce compliance burdens on developers and avoid a patchwork approach" by providing a blueprint for other states, while keeping the public safer. It stops short of advocating for any specific policy, instead outlining the key principles the working group believes California should adopt when crafting future legislation. It "steers clear" of some of the more divisive provisions of SB 1047, like the requirement for a "kill switch" or shutdown mechanism to quickly halt certain AI systems in case of potential harm, says Scott Singer, a visiting scholar in the Technology and International Affairs Program at the Carnegie Endowment for International Peace, and a lead-writer of the report.

Instead, the approach centers around enhancing transparency, for example through legally protecting whistleblowers and establishing incident reporting systems, so that lawmakers and the public have better visibility into AI's progress. The goal is to "reap the benefits of innovation. Let's not set artificial barriers, but at the same time, as we go, let's think about what we're learning about how it is that the technology is behaving," says Cuellar, who co-led the report. The report emphasizes this visibility is crucial not only for public-facing AI applications, but for understanding how systems are tested and deployed inside AI companies, where concerning behaviors might first emerge. "The underlying approach here is one of 'trust but verify,'" Singer says, a concept borrowed from Cold War-era arms control treaties that would involve designing mechanisms to independently check compliance. That's a departure from existing efforts, which hinge on voluntary cooperation from companies, such as the deal between OpenAI and Center for AI Standards and Innovation (formerly the U.S. AI Safety Institute) to conduct pre-deployment tests. It's an approach that acknowledges the "substantial expertise inside industry," Singer says, but "also underscores the importance of methods of independently verifying safety claims."