TSMC has US fast-track export status for China revoked | EU court backs latest data transfer deal agreed by US and EU | NSW government sets up centralised AI office
Why AI labs struggle to stop chatbots talking to teenagers about suicide
Good morning. It's Thursday, 4th of September.
The Daily Cyber & Tech Digest focuses on the topics we work on, including cybersecurity, critical technologies, foreign interference & disinformation.
Follow us on Bluesky, on LinkedIn, and on X.
The U.S. has revoked Taiwan Semiconductor Manufacturing Co's authorization to ship key equipment to its main China facility, the chip manufacturer said on Tuesday. Reuters
A data transfer deal agreed by the European Union and the United States two years ago to replace two previous pacts rejected by a higher tribunal was given the green light by Europe's second-highest court on Wednesday. Reuters
The NSW government has set up a centralised Office for Artificial Intelligence to provide advice and set robust standards to drive uptake of the technology across the state’s public sector. InnovationAus
🚀 We’re rebuilding ASPI’s China Defence Universities Tracker from the ground up. The major expansion adds richer profiles, rankings powered by the Critical Technology Tracker, new mapping of links to China’s state-owned defence industry, analysis of China–Russia research ties, and data on the surge in dual-use research centres—now covering 180+ entities with faster search. Be first to get early-access invites and launch updates: https://unitracker.aspi.org.au/
Australia
NSW government sets up centralised AI office
InnovationAus
Justin Hendry
The NSW government has set up a centralised Office for Artificial Intelligence to provide advice and set robust standards to drive uptake of the technology across the state’s public sector. The new Office for AI, which has been established for an initial two-year period, will sit within the Digital NSW arm of the Department of Customer Service, complementing its existing AI frameworks.
AI-powered road speed and mobile cameras to begin operation in WA
CyberDaily
Daniel Croft
The state has been trialling the technology on its roads during a seven-month introductory period, during which time it issued over 60,000 notices for over 275,000 offences. Six mobile safety cameras, including trailer and fixed, were used during the trial, which began in January, while two fixed camera locations on the Kwinana Freeway were upgraded. The cameras will complement the existing camera fleet.
Aussie super industry to run sector-wide cyber exercise
CyberDaily
David Hollingworth
The superannuation’s Gateway Network Governance Body will take the lead, running a sector-wide cyber security exercise this week, putting Australia’s super industry to the test. “We are pleased to be able to provide these opportunities for all types of organisations across the superannuation ecosystem to come together to explore real-world response strategies in a safe, collaborative setting and exercise our collective response capability,” Michelle Bower, GNGB CEO and exercise director, said in a 2 September statement.
Terrorists are adapting to new drone technology. Australia must follow
The Strategist
Olivia Ortlieb
The advancement and proliferation of autonomous drones as weapons will make terrorism easier, more destructive, and harder to trace—and Australia is underprepared. Modern commercial drones increasingly incorporate automated satellite-navigation targeting and swarming capabilities, bypassing some traditional constraints on terrorism. Mass drone swarm deployment by Ukraine has shown the world the destructive potential of cheap, weaponised drones, and groups such as Islamic State and Hezbollah are striving to adopt similar playbooks.
USA
TSMC, like South Korean rivals, has US fast-track export status for China revoked
Reuters
Karen Freifeld and Wen-Yee Lee
The U.S. has revoked Taiwan Semiconductor Manufacturing Co's authorization to ship key equipment to its main China facility, the chip manufacturer said on Tuesday. The change removes a fast-track export privilege known as Validated End User status, effective Dec. 31, TSMC said, meaning future shipments of American chipmaking tools to TSMC's Nanjing site will require U.S. export licenses. TSMC said it was evaluating the situation and communicating with the U.S. government, adding that it remains "committed to ensuring the uninterrupted operations of TSMC Nanjing."
US curbs TSMC’s tool shipments to China The Wall Street Journal
America's 'New Right' says AI threatens both US and China
Nikkei Asia
Ken Moriyasu
Of the various factions supporting U.S. President Donald Trump in his second administration, two major groups seem increasingly at odds: the "New Right" and tech titans. At the three-day National Conservatism Conference that kicked off here Tuesday, members of the New Right -- mostly young conservatives who blame the establishment for the destruction of the American working class -- distanced themselves from the Silicon Valley types, warning that the pursuit of artificial superintelligence (ASI) threatens to destroy families and jobs and runs counter to the conservative movement.
CISA taps Nicholas Andersen for executive assistant director of cybersecurity
CyberScoop
Tim Starks
Nicholas Andersen is taking over a top leadership role at the Cybersecurity and Infrastructure Security Agency, CISA announced Tuesday. He will become executive assistant director of cybersecurity at the agency in a role that’s seen swift turnover in the past year. It’s a position that has, in the past, led CISA efforts on protecting federal civilian agency networks and protecting critical infrastructure against cyber threats. Andersen is a veteran of the first Trump administration, where from 2019 to 2021 he served in the Department of Energy’s Cybersecurity, Energy Security and Emergency Response division as both the principal deputy assistant secretary and performed the duties of assistant secretary.
Disney agrees to $10 million settlement for collecting data from children
The Record by Recorded Future
Suzanne Smalley
The Federal Trade Commission on Tuesday announced Disney has agreed to pay $10 million to settle allegations that it collected personal data from children watching YouTube videos without parental notification or consent. Following a referral from the FTC, the Department of Justice filed a complaint alleging that Disney broke the Children’s Online Privacy Protection Rule by neglecting to label a “significant number” of videos it places on YouTube as “Made for Kids.”
How disinformation about the Minnesota shooting spread like wildfire on X
WIRED
David Gilbert
Minutes after the perpetrator of the shooting at Annunciation Catholic Church in Minneapolis last week was identified, YouTube appeared to delete several videos they had shared that morning. But not before the videos were downloaded and reshared in full on X. Within hours, the platform was flooded with wild claims about the shooter and her motivation, with everyone from Elon Musk, the site’s owner, to the head of the FBI and left-wing activists posting half-baked allegations blaming anti-Christian hate, transgender genocide, and white supremacy. Many of the posts racked up millions of views per X’s public metrics.
Southeast Asia
Thailand carves out a less-glamorous AI niche: printed circuit boards
Nikkei Asia
Lauly Li and Cheng Ting-Fang
Some 60 kilometers north of Bangkok in the old capital of Ayutthaya, China's Victory Giant Technology is in the hectic final stages of moving new equipment into its newly built printed circuit boar factory. Nvidia's leading supplier of PCB for AI servers and graphic cards is on track to have the plant -- its second in Thailand -- up and running less than a year after acquiring its first one in the Southeast Asian country from Taiwan's APCB Group.
Ukraine – Russia
Moscow reportedly hires hackers who breached city’s school system
The Record by Recorded Future
Daryna Antoniuk
Moscow authorities have hired several hackers who previously launched a cyberattack against the capital’s digital education platform, a city official told state-run media. According to Moscow Deputy Mayor Anastasia Rakova, “three or four young people” who had earlier “successfully managed” to hack the Moscow Electronic School are now working on the educational platform and other city services. She did not disclose their names or provide details of the cyberattack. The platform is used by Moscow’s students, teachers and parents and has faced repeated cyber incidents in recent years.
Europe
EU court backs latest data transfer deal agreed by US and EU
Reuters
Foo Yun Chee and Sudip Kar-Gupta
A data transfer deal agreed by the European Union and the United States two years ago to replace two previous pacts rejected by a higher tribunal was given the green light by Europe's second-highest court on Wednesday. The ruling will provide legal certainty to thousands of companies, ranging from banks to tech companies to drugmakers and car manufacturers, that transfer personal data across the Atlantic for commercial use such as payroll purposes and cloud infrastructure.
EU court dismisses Zalando’s challenge to online content law
The Wall Street Journal
Edith Hancock
The European Union’s general court upheld a decision by tech regulators to impose tougher rules on German online retailer Zalando under the bloc’s digital content law. The court said Wednesday that it backed the decision by the European Commission, the bloc’s executive arm, to label the company—which sells clothing from brands like Mango and Bershka alongside Gucci sunglasses—as a “very large online platform” under the Digital Services Act.
France fines Shein $176 million over cookies
Reuters
Helen Reid
Online fast-fashion retailer Shein received a 150 million euro ($175.61 million) fine on Wednesday from France's data protection authority over the improper use of cookies, a decision the company contested and said it would appeal. The Commission Nationale de l'Informatique et des Libertés, a government body charged with enforcing consumer data protection, said Shein's website failed to comply with regulations in collecting consumers' data without consent.
UK
Jordan escalates global tech argument, with Farage's help
POLITICO
Anthony Adragna and Gabby Miller
A House Judiciary Committee hearing Wednesday pumped fresh oxygen into the transatlantic argument over tech regulation, giving right-wing United Kingdom leader Nigel Farage a platform to attack his own country’s laws. Online safety rules have created an “awful, authoritarian situation” in the European Union and the U.K., said Farage, leader of his country’s Reform Party. Farage was invited by House Judiciary Chair Jim Jordan, marking an escalation in Jordan’s war on Europe’s tech regulations — part of a yearslong effort to pressure Europe into loosening laws that have led to investigations and fines against American tech companies.
Big Tech
Google keeps Chrome and Apple deal but must share data in big antitrust ruling
Reuters
Jody Godoy and Mike Scarcella
Google won't have to sell its Chrome browser, a judge in Washington said on Tuesday, handing a rare win to Big Tech in its battle with U.S. antitrust enforcers, but ordering Google to share data with rivals to open up competition in online search. Google parent Alphabet's shares were up 7.2% in extended trading on Tuesday as investors cheered the judge's ruling, which also allows Google to keep making lucrative payments to Apple that antitrust enforcers said froze out search rivals. Apple shares rose 3%.
Google dodges worst penalties in U.S. antitrust case The Wall Street Journal
Alphabet shares surge after dodging antitrust breakup bullet
Reuters
Akash Sriram and Rashika Singh
Alphabet shares closed more than 9% higher on Wednesday after a U.S. judge ruled against breaking up the Google parent, clearing a major regulatory overhang and adding about $210 billion to the company's market value. The ruling on Tuesday by Judge Amit Mehta allows Google to retain control of its Chrome browser and Android mobile operating system, while barring certain exclusive contracts with device makers and browser developers.
Google's AI rivals get a boost from data-sharing order, but tech giant far from routed
Reuters
Kendrick Kai
A rising group of artificial intelligence companies stand to gain from an antitrust ruling on Tuesday that ordered Alphabet's Google to share its invaluable search data with competitors. Matching Google's heft, though, will take time and huge resources, with no guarantees that any rival product will win as many users, experts said. While Google was spared the devastating outcome of having to sell its popular Chrome browser and Android operating system, U.S. District Judge Amit Mehta's ruling was a nod to regulators' efforts to level the playing field for companies who have invested billions to boost their AI business.
Apple gains on US ruling that spares search deal with Google
Bloomberg
Nick Turner
Apple Inc. shares gained in late trading after a US judge stopped short of barring its lucrative search arrangement with Google, a deal that has generated roughly $20 billion in revenue a year for the iPhone maker. Though Judge Amit Mehta ruled in an antitrust case that Google can’t enter exclusive contracts for internet search, deals that make the search provider a default option in internet browsers are still allowed.
Apple AI researcher for robotics joins Meta in latest exit
Bloomberg
Mark Gurner
Apple Inc.’s lead artificial intelligence researcher for robotics has departed the company to join Meta Platforms Inc.’s competing effort, part of an exodus of AI talent from the iPhone maker. The employee, Jian Zhang, joined the Meta Robotics Studio, the social media company confirmed on Tuesday. Separately, three more AI researchers are leaving Apple’s in-house large language models team, adding to upheaval in that group, according to people with knowledge of the matter. The latest string of departures — all taking place over the last week — includes John Peebles, Nan Du and Zhao Meng, said the people, who asked not to be identified because the moves haven’t been announced.
WhatsApp warns of 'attack against specific targeted users'
The Register
Iain Thomson
A flaw in Meta's WhatsApp app “may have been exploited in a sophisticated attack against specific targeted users.” Meta made that alarming admission last week in a security advisory that disclosed CVE-2025-55177, which it described as allowing “Incomplete authorization of linked device synchronization messages in WhatsApp [which] could have allowed an unrelated user to trigger processing of content from an arbitrary URL on a target’s device.”
WhatsApp, Apple warn of highly targeted attacks with zero-day vulnerability The Record by Recorded Future
Amazon shuts down watering hole attack attributed to Russia’s APT29 hacking group
The Record by Recorded Future
Jonathan Greig
A digital trap set by Russia’s foreign intelligence service was disrupted by Amazon in a recent operation. The company’s threat intelligence team said it identified a so-called watering hole campaign in August where hackers compromised a legitimate website and redirected visitors to malicious infrastructure. Amazon Chief Information Security Officer CJ Moses said the watering hole was the work of APT29 (also tracked as BlueBravo and Cozy Bear), a notorious hacking operation that U.S. officials have long attributed to the Russian Foreign Intelligence Service.
Artificial Intelligence
The problem of AI chatbots discussing suicide with teenagers
Financial Times
Cristina Criddle and Melissa Heikkilä
The world’s top artificial intelligence companies are grappling with the problem of chatbots engaging in conversations about suicide and self-harm, as families claim their products are not doing enough to protect young users. OpenAI and Character.ai are being sued by the parents of dead teenagers, who argue that the companies’ products encouraged and validated suicidal thoughts before the young people took their lives. The lawsuits against groups such as OpenAI underscore the reputational and financial risks for tech companies that have raised billions of dollars in pursuit of AI products that converse with people in a humanlike way.
Fake celebrity chatbots sent risqué messages to teens on top AI app
The Washington Post
Nitasha Tiku
AI-generated chatbots using the names and likenesses of actor Timothée Chalamet, singer Chappell Roan and National Football League quarterback Patrick Mahomes chatted inappropriately with teen accounts on topics including sex, self-harm and drugs, two online safety nonprofit organizations found. The chatbots responded via text and through AI-generated voices trained to sound like the celebrities. The celebrity chatbots were created by Character users with features the app provides to let anyone easily make a custom chatbot, add a synthetic voice and make it available for others to use.
The less you know about AI, the more you are likely to use it
The Wall Street Journal
Heidi Mitchell
When it comes to most new technologies, early adopters tend to be the people who know and understand the tools the best. With artificial intelligence, the opposite seems to be true. This counterintuitive finding comes from new research, which suggests that the people most drawn to AI tend to be those who understand the technology the least. AI seems mysterious and even magical to these people, the researchers found, leading to a sense of awe regarding AI’s ability to complete tasks. This is particularly true when the task is traditionally associated with human attributes, such as writing a poem or creating a new fusion recipe.
The world is at risk of sliding into a new cold war over AI governance
Nikkei Asia
Zhou Xin
As the world's gaze remains riveted on the wars in Gaza and Ukraine, we may be overlooking far more consequential perils unfolding right in our midst. On June 28, China unveiled its Global AI Governance Initiative, a proposal that has slipped under the radar as military conflicts rage. Few people noted that exactly 111 years earlier Archduke Franz Ferdinand was assassinated in Sarajevo, touching off a cascade of events that plunged the globe into an era of extremism.
DeepSeek sheds light on data collection for AI training and warns of ‘hallucination’ risks
South China Morning Post
Vincent Chow
Chinese artificial intelligence start-up DeepSeek has lifted the veil on how it filters data to train its models, raising red flags about “hallucination” and “abuse” risks. In a document published on Monday, the Hangzhou-based start-up said it “has always prioritised AI security” and decided to make its disclosure to help people use its models, at a time when Beijing is ramping up oversight over the industry. The company said data in the pre-training stage was “mainly” collected from publicly available online information as well as authorised third-party data, and DeepSeek had no intention to collect personal data.
AI Startup Flock thinks it can eliminate all crime in America
Forbes
Thomas Brewster
Langley offers a prediction: In less than 10 years, Flock’s cameras, airborne and fixed, will eradicate almost all crime in the U.S. (He acknowledges that programs to boost youth employment and cut recidivism will help.) It sounds like a pipe dream from another AI-can-solve- everything tech bro, but Langley, in the face of a wave of opposition from privacy advocates and Flock’s archrival, the $2.1 billion (2024 revenue) police tech giant Axon Enterprise, is a true believer. He’s convinced that America can and should be a place where everyone feels safe.
Misc
Automated Sextortion spyware takes webcam pics of victims watching porn
WIRED
Andy Greenberg
Sextortion-based hacking, which hijacks a victim's webcam or blackmails them with nudes they're tricked or coerced into sharing, has long represented one of the most disturbing forms of cybercrime. Now one specimen of widely available spyware has turned that relatively manual crime into an automated feature, detecting when the user is browsing pornography on their PC, screenshotting it, and taking a candid photo of the victim through their webcam. On Wednesday, researchers at security firm Proofpoint published their analysis of an open-source variant of “infostealer” malware known as Stealerium that the company has seen used in multiple cybercriminal campaigns since May of this year.
Events & Podcasts
The Sydney Dialogue 2025
The Australian Strategic Policy Institute is pleased to announce the Sydney Dialogue, the world’s premier policy summit for critical, emerging and cyber technologies, will return on 4-5 December. Now in its fourth year, the dialogue attracts the world’s top thinkers, innovators and policymakers, and focusses on the most pressing issues at the intersection of technology and security. TSD has become the place where new partnerships are built among governments, industry and civil society, and where existing partnerships are deepened.
The Daily Cyber & Tech Digest is brought to you by the Cyber, Technology & Security Programs team at ASPI and supported by partners.
