--- date: 2026-08-31 subject: "EU puts ChatGPT under DSA | FSB warns G20 on AI | Abbott cuts Flock funds | Sony sues Anthropic" --- **The European Commission designated ChatGPT a "very large online search engine"** under the EU's content and platform design law, and gave it, along with Reddit and Roblox, until January 2027 to start measuring and curbing systemic risks. **Bank of England pressed G20 finance ministers to set rules** for releasing frontier AI models, writing as chair of the Financial Stability Board that many countries lack protocols to manage systems capable of disrupting global finance. **Texas Gov. Greg Abbott froze state spending on Flock Safety's license plate cameras** amid scrutiny of more than $30 million in purchases and repeated misuse by officers. **Sony Music, Warner Chappell sue Anthropic** and two co-founders over alleged torrenting of copyrighted works. # Top Stories - **EU designates ChatGPT a "very large search engine," setting January compliance date** — The European Commission designated ChatGPT a "very large online search engine" under the Digital Services Act, the EU statute that polices illegal content and platform design. It also named Reddit and Roblox "very large online platforms." All three had themselves reported EU audiences above the 45 million monthly average that triggers designation. Each now has four months from notification, until January 2027, to take on the tier's added duties: measuring and reducing the systemic risks its service and its algorithmic systems create, covering illegal content, harm to minors and interference with elections. [European Commission](https://digital-strategy.ec.europa.eu/en/news/commission-designates-chatgpt-reddit-roblox-under-digital-services-act) - **Bank of England urges G20 finance ministers to write rules on frontier AI releases** — Bank of England Governor Andrew Bailey, who chairs the Financial Stability Board, the global financial risk watchdog, wrote to the G20's finance ministers and central bank governors in a two page letter, per The Guardian. Frontier models are becoming more autonomous, better at solving problems and more capable as threats, he wrote. They could destabilize a tightly linked global financial system through cyber disruption that carries across borders. "Recent developments have also highlighted to me that many jurisdictions do not have the protocols in place to manage the development, release, and deployment of advanced frontier AI models, heightening risks for the financial sector and beyond," he wrote. The letter landed before those ministers convene in North Carolina this week. [The Guardian](https://www.theguardian.com/business/2026/aug/31/advanced-frontier-ai-financial-stability-andrew-bailey-g20) [Financial Stability Board](https://www.fsb.org/2026/08/fsb-chairs-letter-to-g20-finance-ministers-and-central-bank-governors-august-2026/) - **Texas Gov. Abbott freezes state spending on Flock license plate cameras** — Texas Gov. Greg Abbott has frozen state spending on Flock Safety's AI cameras, per The Verge. The freeze came just before a Texas Tribune investigation reported that the state had spent more than $30 million on them. At least six Texas incidents have led to officers being suspended or criminally charged for abusing Flock systems. Gina Hinojosa, the Democrat running against Abbott for governor, has faulted him for expanding the network.[The Verge](https://www.theverge.com/ai-artificial-intelligence/986541/texas-governor-abbott-flock-cameras) - **Sony Music, Warner Chappell sue Anthropic and two co-founders over alleged torrenting of copyrighted works** — Music publishers including Sony Music Publishing and Warner Chappell sued Anthropic in the U.S. District Court for the Northern District of California late Friday, naming co-founders Dario Amodei and Benjamin Mann as defendants and alleging a "brazen campaign of illegally torrenting, scraping, and downloading copyrighted works," TechCrunch reported. The complaint accuses the company of using thousands of copyrighted works to train Claude. Lawyers on the case also act for Concord Music Group and Universal Music Group in a January complaint against Anthropic. They led *Bartz v. Anthropic*, where a judge ruled it was lawful for the lab to train on copyrighted works but not to obtain them through piracy. Anthropic agreed to pay $1.5 billion to settle that case. "We disagree with the publishers' claims and we intend to defend ourselves robustly in court," an Anthropic spokesperson said. [TechCrunch](https://techcrunch.com/2026/08/29/sony-music-warner-sue-anthropic-alleging-a-brazen-campaign-of-intellectual-property-theft/) - **OPM tells agencies common AI uses in federal hiring escape high impact testing rules** — Office of Personnel Management Director Scott Kupor advised department heads Thursday that several common uses of AI in federal hiring are not high impact under federal AI governance policy, FedScoop reported. His memo puts AI that writes job announcements, evaluates applicants, reviews files before an offer and tracks hiring metrics outside that category, so agencies need not run pre-deployment testing, impact assessments or monitoring for adverse effects. Federal employment determinations are among the uses that the Office of Management and Budget's AI governance memo presumes to be high impact. OPM said agencies can rebut that presumption when an official independently reviews the data and adopts the AI assisted product. "Too often, agencies were defaulting to the view that any use of AI in hiring was subject to the high-impact standard," OPM Chief Information Officer Adam Starr wrote. [FedScoop](https://fedscoop.com/opm-memo-federal-hiring-ai-usage/) - **Climate movement organizers launch Irreplaceable to press for slowing the AI race** — An advocacy group called Irreplaceable, named for the argument that people cannot be replaced by machines, formally launched Aug. 28, The Atlantic reported. Its staff comes from the climate movement, and it pitches its campaigning at Americans' emotional reaction to AI rather than at white papers and policy recommendations. Executive Director Phil Aroneanu, a co-founder of 350.org, said climate advocates initially floundered because they behaved like policy wonks. Campaign strategist Jeremy Ornstein, formerly of the Sunrise Movement, said: "People have the intuition that if we create machines more powerful than people, that could go dangerously and terribly wrong." In polling, roughly three-quarters of Americans oppose a data center in their area and nearly 70% say the technology is advancing too quickly. [The Atlantic](https://www.theatlantic.com/technology/2026/08/irreplaceable-climate-activists-ai-backlash/688404/) # China Watch - **Chinese court freezes Nexperia assets as Wingtech presses claim over Dutch seizure** — Wingtech Technology said the Dongguan Intermediate People's Court in Guangdong province ruled Aug. 28 to impound and freeze assets worth 2.14 billion yuan (about $320 million) belonging to Nexperia Ltd. and ITEC B.V., per Jiemian. The order took effect immediately, and the court applied further preservation measures the same day at the request of Wingtech and its affiliate Yucheng Holdings. Wingtech bought Nexperia in 2019 and lost control of the chipmaker in October 2025, when the Dutch government intervened. It sued in Dongguan on May 23 under China's Anti-Foreign Sanctions Law, asking the court to declare the seizure unlawful, return control of core assets and award damages. [Jiemian](https://www.jiemian.com/article/15036356.html) - **Uzbekistan asks to join U.S. AI supply chain bloc a month after signing on to China's** — Uzbekistan has formally asked to join Pax Silica, the U.S. framework for securing AI supply chains, its ambassador to Washington, Furqat Sidikov, told the South China Morning Post. Sidikov said Tashkent sent the State Department a diplomatic note to that effect last week. Uzbekistan, Kazakhstan and Russia are among the 29 founding members of the World Artificial Intelligence Cooperation Organization, the competing body China set up in Shanghai in July. The State Department has drafted a notice telling countries that want wider access to advanced U.S. technology under Pax Silica to stay out of the Chinese group. Pax Silica, launched in late 2025, covers about two dozen economies plus the EU. [South China Morning Post](https://www.scmp.com/news/china/diplomacy/article/3365645/uzbekistan-tests-us-china-ai-rift-bid-join-rival-tech-bloc) - **China's CXMT starts volume output of newest phone memory standard with Xiaomi first in line** — ChangXin Memory Technologies, or CXMT, has started mass production of LPDDR6, the latest standard for the working memory in smartphones, the company and Xiaomi said Aug. 29, per TechNode. They named Xiaomi's 18 Fold as the first phone to use it. Xiaomi has confirmed the foldable will go on sale in September. Xiaomi's own Xring O3 processor, unveiled Aug. 24, is built to work with the standard, putting a domestically designed processor and domestically made memory in the same handset. CXMT, China's largest memory chipmaker, reached volume output of the previous generation less than a year ago. [TechNode](https://technode.com/2026/08/31/cxmt-starts-lpddr6-mass-production-for-xiaomis-upcoming-18-fold/) - **SenseTime chip spinout Sunrise nearly doubles valuation in new inference chip round** — Sunrise, the AI chipmaker carved out of SenseTime's chip division, has raised 2 billion yuan (about $300 million) in a round that values it at about 20 billion yuan ($3 billion) after the investment, per TechNode, citing an Aug. 28 report. That is close to twice what investors put on the company in its previous round. Sunrise designs graphics processors aimed only at inference, the work of running a finished model rather than training one. Chinese companies are pushing domestic alternatives in that segment for deploying large models at scale. Sunrise was set up inside SenseTime in 2020, became an independent company in 2024 and raised more than $140 million as recently as April. [TechNode](https://technode.com/2026/08/31/sensetime-spinout-sunrise-reportedly-raises-rmb2-billion-at-rmb20-billion-valuation/) - **State affiliated Chinese paper urges human check on AI translations of official language** — China should expand its translation capacity and screen for the distortions AI introduces, a commentary in a state affiliated academic paper said, warning that literal machine renderings of official wording can make Beijing's intentions read as more menacing abroad, per the South China Morning Post. Li Yan, a professor at Xi'an Fanyi University, said human translators should work alongside AI systems on political terminology to catch cultural misreadings, gaps in logic and value laden bias. Chinese state media has separately pressed readers to drop English AI jargon in favor of Chinese terms. [South China Morning Post](https://www.scmp.com/news/china/diplomacy/article/3365635/ai-can-make-chinas-political-language-sound-more-threatening-scholar-warns) - **Chinese makers behind 86% of humanoid robot shipments; FCC Covered List now reaches robots** — Counterpoint Research put global humanoid robot shipments at 22,000 units in the first half of this year, with the five largest makers, all Chinese, accounting for 86% of them, per TechCrunch. Counterpoint analyst Soumen Mandal said U.S. companies operate at far smaller scale and that Chinese makers are cutting costs by building more of the technology stack themselves. Over July and August, Washington curbed the use of advanced robotic systems built overseas and set steep tariffs on imported drones and drone components, citing national security for both. The drone tariffs take effect in September, with further component tariffs in 2027. The FCC's Covered List, the roster of gear barred from U.S. networks, began in 2021 with telecom and surveillance equipment from Huawei, ZTE and Hikvision. It now reaches advanced robotic devices. [TechCrunch](https://techcrunch.com/2026/08/30/the-u-s-is-building-barriers-around-drones-and-robots-china-still-has-scale/) # Policy Tracker - **[Exec] VA puts large language model chatbot on its help page in beta** — The Department of Veterans Affairs launched an AI helpdesk chatbot driven by a large language model, FedScoop reported Aug. 28. The bot is in beta testing and sits on VA's "Contact Us" page. It cannot hand a user off to a live person. A disclaimer on the page warns against using it for emergencies or medical questions and says its answers can be wrong. Luciana Morais, a product owner at VA, wrote on LinkedIn that reaching launch included putting the department's own AI governance documents in place. Sign-in personalization, live agent support and use across more VA pages are planned later this year. [FedScoop](https://fedscoop.com/va-launches-ai-helpdesk-chatbot/) - **[Exec] ICE forecasts up to $2 million for Boston Dynamics robot dogs** — Immigration and Customs Enforcement plans to spend between $1 million and $2 million on Boston Dynamics' Spot robots, according to a Homeland Security funding document posted Aug. 27, per The Independent. The four-legged machines would help agents inspect sites, keep track of their surroundings and gauge hazards where sending people would be risky. Spot weighs roughly 75 pounds, carries cameras and Wi-Fi and travels at up to 5 feet per second. It can be driven remotely or run assigned missions on its own, per the Spanish news agency EFE. The document sits on DHS's Acquisition Planning Forecast System, which lists intended purchases rather than awards. [The Independent](https://www.the-independent.com/news/world/americas/us-politics/ice-robot-dogs-trump-immigration-b3041608.html) [EFE](https://efe.com/mundo/2026-08-29/estados-unidos-inversion-perros-robots-operativos-ice/) [The Hill](https://thehill.com/homenews/administration/6058739-ice-robot-dogs-2-million-plans/) - **[State] North Dakota legislative panel takes public testimony on AI and data centers** — North Dakota's Artificial Intelligence and Data Center Committee convenes Monday at 9 a.m. at the Capitol in Bismarck, with public comment scheduled for 11:15 a.m., per KFGO. Policy experts, the energy industry and the state's utility and pollution regulators are all scheduled to present. The panel, announced in June, has already taken testimony on AI regulation, AI use in education, risks to young people and objections to data centers. Multiple data center projects are proposed in the state, with developers drawn by its cool climate, access to power and available land. [KFGO](https://kfgo.com/2026/08/29/public-can-testify-monday-at-ai-and-data-center-committee/) # Capability & Research Watch - **Anthropic paper reports automated researchers beat human proposals on alignment fixes** — Anthropic published a paper Aug. 28 titled "Automated Researchers Can Reliably Mitigate Alignment Failures," led by Anthropic fellow Chen Yueh-Han, TechCrunch reported. Across 10 benchmarks that each target a specific misaligned behavior, the automated systems raised scores on all of them while leaving overall performance intact. Each automated alignment researcher, or AAR, searches the literature, proposes a method, trains the model on it for 30 minutes and iterates, keeping the methods that work. "The best AAR method beats what experienced humans propose, on average within six hours," the paper says. It puts the cost at roughly $4 an hour in model usage against $150 an hour for its human researchers. The authors note the approach holds only insofar as the benchmarks reflect the actual alignment goals. [TechCrunch](https://techcrunch.com/2026/08/28/an-anthropic-researcher-just-gave-us-a-peek-at-self-improving-ai/) [Anthropic](https://www.anthropic.com/research/automated-researchers-mitigate-alignment-failures) - **NHS patient watchdog documents AI scribes misnaming drugs and diagnoses** — Healthwatch England, the statutory patient champion for the National Health Service, warned that AI tools transcribing consultations are recording the wrong drug and disease names, The Guardian reported. In one case a scribe's summary recorded a woman as having demyelination, a serious form of nerve damage associated with multiple sclerosis, when her MRI result should have been entered as "null demyelination." The hospital fixed it only after the patient, herself an NHS health professional, questioned the tool's account. In another, a scribe swapped the drug a general practitioner had prescribed for a similarly named one, again caught by the patient rather than the doctor. Healthwatch said it has heard "multiple stories from patients who have noticed these errors when a health professional hasn't." England's 10 year NHS plan expects AI scribes to cut staff paperwork. [The Guardian](https://www.theguardian.com/society/2026/aug/31/doctors-ai-scribes-get-names-of-drugs-and-diagnoses-wrong-nhs-watchdog-warns) # Industry & Market Watch - **Musk confirms SpaceX foundry to cast gas turbine blades for data center power** — Elon Musk confirmed on X on Saturday that a secretive SpaceX foundry near Bastrop, Texas, will cast blades and vanes for gas turbines, TechCrunch reported. Casting those parts is the bottleneck in turbine production, he wrote. Doing it in-house could bring new gas turbines online as much as 18 months sooner. GE Vernova, which builds the turbines, says AI infrastructure demand has taken up nearly all of its capacity through 2030. The International Energy Agency expects worldwide data center power consumption to about double by then. Due diligence analyst Corey Trinetti reported that SpaceX bought roughly 830 acres near its Bastrop Starlink factory between March and June.[TechCrunch](https://techcrunch.com/2026/08/30/musks-faster-path-to-more-gas-turbines-comes-with-pollution-problem/) - **SB Energy draft IPO documents value OpenAI's tenant warrants at $5.5 billion** — OpenAI holds warrants in SB Energy now valued at an estimated $5.5 billion, awarded as part of the arrangement that secured OpenAI as a data center tenant, The Wall Street Journal reported from draft IPO documents. The warrants were worth $3.6 billion when granted in January. OpenAI put another $500 million into the company earlier this year. After the offering, which could raise $5 billion to $7 billion as soon as next month, OpenAI is expected to hold a single digit stake. SB Energy, controlled by SoftBank, has nearly nine gigawatts of contracted computing capacity and no operating data centers. Its data center unit expects a contracted backlog above $400 billion. SB Energy flagged its heavy reliance on OpenAI as a risk to investors. [Investing.com](https://www.investing.com/news/stock-market-news/sb-energy-offered-openai-55-bln-in-warrants-to-secure-datacenter-deal--wsj-4882171) - **Incumbents buy up the open-weight AI stack as Nvidia's Hugging Face deal stays unconfirmed** — Nvidia's reported $13 billion acquisition of Hugging Face, the main repository where developers share open-weight AI models, remained unconfirmed on Aug. 28, TechCrunch reported. That is above the $12.9 billion figure reported by The Information in AIPD's [August 27th edition](https://aipolicydaily.org/archive/daily/2026-08-27/). Those talks follow a $6 billion agreement with model builder Poolside that moves most of its staff to Nvidia. Two weeks earlier, Stripe bought OpenRouter, which routes open-weight models to businesses, for more than $7 billion. Nvidia already ships its own open-weight Nemotron models, without large uptake. Just 6% of companies use open-weight models, per spending data from Ramp, and inference costs are pushing firms toward cheaper Chinese models from Moonshot, DeepSeek and Alibaba. [TechCrunch](https://techcrunch.com/2026/08/28/open-weight-ai-companies-are-the-valleys-hottest-acquisition-targets/) - **Nvidia's robot chip business sells into China under rules barring its top AI chips** — Nvidia's physical AI line, the chips and software it sells to builders of robots, cars and robotaxis, generates about $10 billion in annual revenue, per The Wall Street Journal. That is a small slice of the $303 billion Nvidia took in over the four quarters ended in July. Chief Executive Jensen Huang said the line will grow tenfold within a decade. U.S. rules block the company's most powerful AI chips from China while leaving its robotics and automotive chips unrestricted. China is the world's largest industrial robot market. Nvidia pairs those chips with developer software and has released open-weight robot models named GR00T and Cosmos. Counterpoint Research analyst Brady Wang said China is central to Nvidia's physical AI road map. [LiveMint](https://www.livemint.com/companies/nvidia-wants-to-run-the-world-s-robots-china-is-an-eager-customer-11788054464203.html) # Global & Geopolitics - **Japan seeks $55.6 billion defense budget centered on attack drones and AI command tools** — Japan's Defense Ministry disclosed a defense budget request of 8.9 trillion yen ($55.6 billion) for the fiscal year ahead, built around drones, AI and what it calls new methods of warfare, AP reported. The preliminary figure is a small increase over the 8.8 trillion yen ($55 billion) requested for the current year. It will rise once Prime Minister Sanae Takaichi's government approves a new security and defense strategy later this year. The ministry plans an integrated AI platform for command and control to speed battlefield assessment and target selection, plus a next generation "AI orchestrator" that would combine multiple AI systems and select among them autonomously. Officials said Japan is weighing whether to use existing foreign AI, including American systems, alongside or combined with domestic development. [AP](https://apnews.com/article/japan-defense-budget-drones-ai-bdda63e2831ff47a214d786bd4e45907) - **India issues $13.4 billion chip scheme covering design, packaging and equipment** — India's Ministry of Electronics and Information Technology issued the framework for Semicon 2.0, setting fiscal support across six segments of the chip industry from design and fabrication to advanced packaging, equipment and talent, per Moneycontrol. Silicon wafer fabrication plants get 40% fiscal support and legacy packaging 25%. The scheme carries an outlay of about 1.3 trillion rupees ($13.4 billion). The government projects roughly 4 trillion rupees ($42 billion) in investment and 2 trillion rupees ($21 billion) in production over its life. Projects run up to six years, and the application window stays open for three. The India Semiconductor Mission will implement the scheme. [Bloomberg](https://www.bloomberg.com/news/articles/2026-08-31/india-offers-chip-design-gear-subsidy-from-13-4-billion-fund) [Moneycontrol](https://www.moneycontrol.com/news/business/companies/semicon-2-0-guidelines-notified-govt-targets-full-semiconductor-value-chain-advanced-packaging-14018825.html) - **UK government rejects Green Party call for data center moratorium** — On Aug. 28, the UK government rejected Green Party leader Zack Polanski's call to "slam the brakes on these energy-guzzling, water-guzzling datacentres," The Guardian reported. The government said the moratorium, which would exempt only local scale facilities serving local communities, "would be a disaster for jobs and national security." A spokesperson said a pause would push investment and jobs overseas and force Britain to rely on foreign computing to handle sensitive health and defense records. Labour has designated data centers as critical national infrastructure, a status Polanski called outrageous. The spokesperson said the designation does not give the facilities priority access to water or electricity. Data centers take about 2.5% of the UK's electricity supply, a share predicted to rise fourfold by 2030. [The Guardian](https://www.theguardian.com/uk-news/2026/aug/28/labour-green-party-zack-polanski-ai-datacentres)