US export control on Anthropic’s Claude Fable 5 and Mythos 5

Plus, ASPI Report: the need to redesign intelligence product for the AI age

Welcome to the latest edition of ASPI’s Cyber & Tech Digest.

Each week, ASPI curates and contextualises the most important developments in cyber, technology, and geopolitics — highlighting what matters and why.

This edition covers the period: 13 June 2026 to 19 June 2026.

Follow the Australian Strategic Policy Institute on Bluesky, LinkedIn, and X.

---

What We’re Tracking

US export control on Anthropic’s Claude Fable 5 and Mythos 5

What happened: On 12 June, two days after Anthropic made Claude Fable 5 public, the US government issued an export-control directive requiring it to suspend access to both Fable 5 and Claude Mythos 5 for foreign nationals. Washington has not publicly explained the directive in detail, but Anthropic has suggested officials were concerned that users could circumvent model safeguards for malicious purposes.

Anthropic disabled the models globally, arguing it could not reliably verify nationality or selectively restrict access at scale. International reaction has focused less on Fable itself than on the precedent: the US can abruptly restrict access to frontier AI systems, including for allies, firms and researchers.

Many cybersecurity experts have questioned whether the ban is proportionate. They argue that while Mythos is particularly strong at bug-finding and exploit generation, Fable was designed with stronger guardrails, and other leading US and Chinese models can perform similar cybersecurity tasks. At the G7, leaders discussed a ‘trusted partners’ framework for allied access to advanced US AI models, particularly for cybersecurity. Meanwhile, more than 100 cybersecurity leaders and experts, led by Alex Stamos, a leading a cyber expert and chief product officer at AI start-up Corridor, have urged Washington to lift the ban, arguing that the models have legitimate defensive uses and that cutting off access could weaken cyber defenders.

Why we’re tracking this: The decision is significant because it appears to be the first time Washington has applied export-control logic to frontier AI model access itself, rather than only to chips, compute or infrastructure. For US allies, including Australia, the issue is reliability: if frontier AI capabilities can be withdrawn overnight under US national security authorities, governments and firms will need to treat AI dependence as a strategic risk. European and UK political reactions have already framed the episode as a warning about technological dependence and AI sovereignty.

What people are saying:

• ‘We are in a race against our adversaries to find and fix these bugs. We are in a race against our adversaries to build defenses.’ — Alex Stamos, chief product officer at AI start-up Corridor and lead author of open letter representing over 100 cybersecurity leaders and experts

• ‘America needs a regulatory system that mitigates AI’s risks, while facilitating its benefits – not one that enables the president to kneecap his least-favorite companies on dubious grounds.’ — Eric Levitz, Vox

• ‘Whatever the machinations in Washington, the latest move fits a trend towards greater national control of the most powerful technology the world has ever seen. Australia must do all it can to make sure we are not left behind.’ — David Wroe, ASPI Senior Fellow

My view: The export-control directive to Anthropic marks an important moment in frontier AI governance. It does not necessarily prove that fully public access to advanced AI is ending, especially when similar capabilities may already exist across other US and Chinese models. But it does show that governments are increasingly willing to restrict access to frontier systems when they believe those systems have national-security implications.

The issue is bigger than Anthropic’s relationship with Washington. It reflects how strategic competition is reshaping the way governments deal with technology companies, especially as AI becomes more powerful, unpredictable and strategically consequential. Frontier models are increasingly being treated not as ordinary software, but as dual-use capabilities whose distribution may be shaped by export controls, alliances and national-security considerations.

One possible outcome is a more segmented AI market: capable but constrained public models; more powerful systems available through trusted-access arrangements; and sovereign or classified models reserved for national-security use. For other governments, the lesson is clear: dependence on foreign AI systems carries strategic risk. If access to frontier models can be restricted abruptly by another country’s national-security decisions, states will need stronger domestic capabilities, diversified partnerships and clearer arrangements with trusted allies.

— Dr Gatra Priyandita, CTS

---

---

What We’re Watching

A weekly scan of notable developments we’re tracking across technology, policy, and geopolitics.

⸻

🚀 Strategic competition

The Trump administration imposed export controls on Anthropic’s Fable 5 and Mythos 5 AI models, prohibiting access by foreign governments, companies, individuals and foreign nationals working inside the United States.

White House adviser David Sacks said on X that a trusted testing partner had identified a jailbreak capable of bypassing Fable’s safeguards and exposing Mythos-level cyber capabilities, and that administration officials had asked Anthropic to fix the vulnerability or withdraw the model before resorting to the export order.

Anthropic responded by disabling access to both models globally, citing the broad scope of the restrictions, but disputed the characterisation of the security issue and argued the cited vulnerability was minor and comparable to those found in other publicly available frontier models.

Politico reported that the controls followed urgent discussions with CEO Dario Amodei and that concerns raised by Amazon contributed to the decision. Senior Anthropic technical staff subsequently travelled to Washington to meet White House officials and held virtual meetings with administration counterparts in an effort to resolve the dispute.

According to Politico, Trump officials and Anthropic have since met to discuss a resolution, with the disagreement centred on whether the Fable vulnerability was sufficiently serious to warrant the action taken.

WIRED reported that administration officials are pressing Anthropic to address all jailbreak vulnerabilities, while cybersecurity experts say completely preventing jailbreaks may be technically infeasible. The administration reportedly does not intend to grant exemptions for G7 allies, and the restrictions are creating uncertainty around Anthropic’s AI releases and planned IPO.

The controls have drawn responses from multiple governments. Canadian Prime Minister Mark Carney said the ban highlights the risks of dependence on a small number of advanced AI systems and compared model concentration risks to systemic vulnerabilities exposed during the 2008 financial crisis.

The UK government denied reports that it had sought an exemption from the restrictions while confirming it remains in contact with both the US government and Anthropic.

The European Commission said it is assessing practical implications of the controls and noted that any contingency measures should not discriminate against partners. European policymakers, according to Politico, used the episode to renew calls for accelerating domestic AI model development and expanding computing capacity, including support for Mistral.

In The Australian Financial Review, the Anthropic export control episode was cited as a warning for Australia to strengthen its position in the AI ecosystem, leverage alliance relationships and address regulatory barriers that may discourage AI companies from training models domestically.

The episode also reignited debate in India over dependence on foreign AI providers and the need for domestic AI capabilities.

China’s commerce ministry criticised the US Defense Department’s decision to add major Chinese technology companies — including Alibaba, Baidu, BYD, NIO, Trina Solar and JA Solar Technology — to its list of firms alleged to support China’s military. Beijing said the move undermined recent understandings reached between US and Chinese leaders and warned of a response if Chinese companies were not treated fairly.

Geoff Wade, in his ASPI’s The Strategist article, argued that more than 200 overseas organisations promoting ‘peaceful reunification’ with Taiwan operate as part of a global Chinese Communist Party influence network linked to the United Front Work Department, disseminating CCP messaging through local Chinese diaspora organisations and Chinese-language media, including in Australia.

China is simultaneously accelerating the integration of AI across military systems, including electronic warfare, communications, autonomous drones and decision-support tools. Analysts say the People’s Liberation Army is pursuing an ‘intelligentized warfare’ strategy combining AI, robotics and unmanned systems, with China aiming to become a global AI leader by 2030, though experts say the extent of current deployment remains difficult to assess.

In The Strategist, John Coyne argued that China is increasingly treating supply-chain information as a strategic asset through new regulations governing industrial and supply-chain security, giving Chinese authorities greater scope to scrutinise information-gathering activities and limit cooperation with certain foreign investigations.

France and Canada signed a General Security of Information Agreement establishing a framework for the secure exchange and protection of classified information, intended to support deeper cooperation in defence, intelligence, space, cybersecurity and artificial intelligence.

⸻

🧠 AI models, agents & compute

US-based developers and smaller technology firms are adopting Chinese models such as DeepSeek, Minimax, Kimi and Xiaomi MiMo for coding, automation, voice recognition and business workflows, often at a fraction of the cost of comparable US services, and many are accessing them through American cloud providers or self-hosted deployments.

Microsoft is evaluating a Microsoft-hosted fine-tuned version of DeepSeek V4 or another open-source model as a lower-cost option for its Copilot Cowork enterprise AI tool, as part of a broader shift toward a multi-model strategy.

DeepSeek reportedly required prospective investors in its first external fundraising round to agree not to recruit its employees or encourage them to establish competing ventures, following the loss of several key researchers to rivals including Xiaomi and Tencent. The company is said to have completed a round valuing it at more than $50 billion.

A former Hugging Face Asia-Pacific ecosystem lead argued in Rest of World that open-source AI development is more collaborative than competitive, with Chinese and US labs mutually benefiting from each other’s research, though he warned that some Chinese labs may reduce open-source releases if sustainable revenue models are not found.

Microsoft CEO Satya Nadella argued on X that organisations will need to build both ‘human capital’ and ‘token capital’ in the AI era, with competitive advantage coming from AI learning loops, private evaluation systems and institutional knowledge bases rather than reliance on frontier models alone. Nadella also warned against excessive concentration of economic value in a small number of AI models, contending that a sustainable AI economy requires broad-based participation.

Anthropic released a major update to Claude Design, adding enterprise-focused design system imports, bidirectional integration with Claude Code, expanded export options and lower token consumption, as part of a broader strategy to embed Claude across enterprise workflows including integrations with Adobe, Canva, Replit and Vercel.

SpaceX’s IPO filing highlights the growing role of Gulf sovereign wealth funds in financing leading AI and technology companies, with investments and partnerships involving Saudi Arabia’s Public Investment Fund, the UAE’s MGX and G42, and major AI firms including OpenAI, Anthropic, xAI and SpaceX. The filing also reveals the scale of long-term Gulf investment in Elon Musk’s businesses and ties Gulf capital to commitments to build AI infrastructure, including data centres, within Saudi Arabia and the UAE.

Australia’s rapid expansion of AI-related data centres is increasingly targeting highly productive agricultural land, particularly in New South Wales and Victoria. Analysis by Digital Agriculture Services found that approximately $21 billion worth of farmland, covering more than 2.1 million hectares and 21,000 farms, lies within zones attractive to data centre developers.

A new report by the Tech Policy Design Institute found Australia has stronger AI capabilities than commonly perceived, particularly in critical minerals, strategic data assets, data centres and international norm-shaping, while identifying gaps in regulatory oversight, public-interest computing and data governance.

Senior Trump administration officials have discussed possible mechanisms for the US government to receive equity stakes in major AI companies. Treasury Secretary Scott Bessent reportedly favoured directing any equity holdings to Trump Accounts, while Commerce Secretary Howard Lutnick preferred a sovereign wealth fund model. The discussions remain preliminary with no formal proposal presented to industry.

The AI data-centre boom is driving demand for advanced memory, wiring, power systems and cooling equipment from South Korean and Taiwanese companies including SK Hynix, Samsung, Foxconn and Delta Electronics.

Japan’s Ajinomoto is a key supplier of insulating film essential for AI chip packaging, with rising demand creating bottlenecks in critical materials for semiconductor manufacturing.

⸻

🛡 Cyber posture

The FBI, working with Google and Black Lotus Labs, disrupted a China-based phishing-as-a-service operation known as Outsider Enterprise, which used AI-assisted phishing kits and more than one million fraudulent URLs to steal credentials and payment data, with authorities estimating contributions to the theft of more than 3.8 million credit card records and approximately US$1.9 billion in losses.

Google separately filed a lawsuit against Outsider Enterprise, alleging the group used Gemini AI to create hundreds of fake corporate and government websites for financial scams targeting Americans, and said it is coordinating with the FBI and major US telecommunications providers to disrupt the network.

The FBI revealed a 22,000-square-foot Cyber Range in Huntsville, Alabama, designed to simulate cyberattacks against realistic infrastructure including connected homes, a hospital, gas station and a data centre with more than 200 servers. All systems are isolated from external networks to prevent malicious code from escaping.

A Chinese-linked cyberespionage group known as UNC6508 conducted a long-running campaign against US and Canadian academic, medical and military research institutions between September 2023 and November 2025, according to Google. The hackers targeted information related to defence intelligence, Indo-Pacific military strategy, artificial intelligence, unmanned systems, cyber warfare and medical research, exploiting vulnerabilities in REDCap servers and automating the exfiltration of emails containing selected keywords.

Researchers from Deep Specter Research say GitHub rejected two vulnerability reports describing design features now being exploited by variants of the Shai-Hulud supply-chain worm. The researchers argue the worm abuses GitHub’s handling of commit timestamps and author metadata to disguise malicious code as legitimate historical contributions, and identified at least 516 malicious software packages, more than 3,000 affected repositories and over 200 compromised developer accounts linked to the campaign.

Senator Mark Warner warned that staffing reductions, leadership vacancies and budget cuts at the Cybersecurity and Infrastructure Security Agency could weaken US cyber defence, citing the loss of roughly one-third of the agency’s workforce and reduced support to state and local governments.

A Senate Armed Services Committee amendment to create a standalone US Cyber Force was narrowly defeated 14–13 during consideration of the fiscal 2027 National Defense Authorization Act, with the bill instead advancing a major reorganisation of Pentagon cyber governance through a new Under Secretary of Defense for Cyber, Information, and Networks.

The UK’s National Cyber Security Centre reported that hostile states are responsible for roughly 75% of cyber incidents affecting Britain’s critical national infrastructure, with NCSC chief Richard Horne disclosing the agency handled more than 200 infrastructure-related incidents in the year to May.

Finland brought charges against the captain and bosun of the cargo vessel Fitburg for damaging two subsea telecommunications cables and attempting to damage eight other connections in the Baltic Sea, with the vessel alleged to have dragged a damaged anchor along the seabed for at least 130 kilometres.

Recorded Future identified more than 36 inauthentic websites linked to sanctions evasion networks supporting Iranian and Russian shadow fleet operations, with the infrastructure impersonating ship registries, maritime administrations and insurance providers to generate fraudulent maritime documentation.

A Russian-speaking ransomware group, The Gentlemen, has claimed responsibility for a cyberattack that disrupted operations at Mackay Sugar’s Racecourse and Farleigh mills in Queensland, halting harvesting across many supplier farms. Mackay Sugar confirmed that an external party accessed parts of its IT environment, while the Gentlemen group has threatened to publish allegedly stolen data if a ransom is not paid.

Russian software company Kaluga Astral, meanwhile, disclosed a cyberattack that disrupted its tax reporting and electronic document management services for about a week, affecting businesses and government-linked organisations.

The Belarus-linked GhostWriter hacking group has expanded its phishing operations from work accounts to personal Gmail accounts belonging to Polish public figures, government officials, researchers, journalists, law enforcement personnel and their relatives, according to CERT Polska. The group has used a growing number of phishing domains to steal credentials and two-factor authentication codes.

A study by the Institute of the Estonian Language found that four versions of Mistral AI performed poorly at identifying or filtering Kremlin narratives compared with other leading AI models, with Mistral models scoring below 40% in detecting Russian propaganda and at times citing sanctioned pro-Russian sources. The findings add to earlier assessments linking Mistral outputs to content amplified by the Pravda network, a large ecosystem of pro-Russian websites.

⸻

🕵️ Surveillance states

Russia’s intensified internet controls have driven many users to adopt workarounds such as VPNs, multiple phones and parallel messaging platforms to access blocked foreign services. The restrictions have disrupted banking, transport, e-commerce and communications, prompting public frustration ahead of Russia’s September parliamentary election and leading authorities to soften some aspects of their messaging on internet controls.

France’s disinformation monitoring agency Viginum accused Israeli technology company BlackCore of conducting online influence operations targeting Scottish First Minister John Swinney, the Scottish National Party and the Scottish Government during Scotland’s 2025 parliamentary election campaign. Viginum said the activity involved coordinated use of at least 256 social media accounts that generated approximately 1,400 comments, but said investigators were unable to identify the client or sponsor behind the operation.

A data leak exposed information linked to Dialog, a private invitation-only network founded by Peter Thiel and Auren Hoffman. The leaked data includes names and personal details of more than 100 people connected to the organisation, as well as information about an upcoming August 2026 gathering in Dublin and discussion topics planned for the event.

⸻

⚖️ Platform accountability

Research by British Future found that X declined to remove many posts containing racial slurs directed at prominent UK politicians when they were reported as hate or harassment, though the platform acted more consistently when content was reported as potentially illegal under the UK Online Safety Act.

Niantic Spatial confirmed that AI models trained using location scans voluntarily submitted by Pokémon Go players are being used in a partnership with drone software company Vantor, aimed at improving drone navigation in GPS-denied environments with potential military applications. Niantic Spatial disputed reports that Pokémon Go data is being directly shared with Vantor, saying player-generated AR scans were used only to train foundation models rather than being directly transferred and that sharing such data is not part of its agreement with Vantor.

A Dutch court artist received damages after a politician from the far-right Party for Freedom used one of her courtroom sketches without permission and altered it with AI before publishing it on social media. Following legal action, the politician apologised, accepted responsibility, paid damages and removed the video.

Tesla presented self-published safety statistics to European regulators in support of approval for its Full Self-Driving (FSD) system, but independent researchers said the figures relied on flawed comparisons that overstated FSD’s safety performance. Dutch regulators approved FSD in April and are pursuing EU-wide approval, while regulators in Sweden, Norway and Greece have also considered or referenced Tesla’s safety claims.

⸻

🧑‍⚖️ Courts, enforcement & regulation

A coalition of US state attorneys general has launched an investigation into OpenAI, with New York issuing a subpoena seeking documents related to advertising, user engagement, consumer and health data practices, impacts on minors and seniors, and model behaviour — coming shortly after OpenAI confidentially filed for an initial public offering.

A US federal judge dismissed with prejudice a lawsuit by Elon Musk’s xAI alleging OpenAI misappropriated trade secrets related to the Grok chatbot, finding no evidence that OpenAI induced a former xAI engineer to disclose confidential information.

South Korea’s Personal Information Protection Commission fined Coupang and its logistics subsidiary 624.7 billion won (US$409 million) following a breach affecting more than 33 million customer accounts, concluding that the breach resulted from basic security failures including the theft of authentication credentials by a former employee.

Ukrainian national Oleksii Oleksiyovych Lytvynenko pleaded guilty in the United States to conspiracy to commit wire fraud for his role in the Conti ransomware operation between 2021 and 2022, with Conti having targeted more than 1,000 victims and collected over $150 million in ransom payments before its collapse.

The European Parliament voted to lift the immunity of Italian MEP Fulvio Martusciello, allowing Belgian authorities to investigate bribery allegations linked to Chinese technology company Huawei as part of a broader Belgian probe launched in 2025 into alleged corruption and money laundering connected to activities in the Parliament.

Section 702 of the Foreign Intelligence Surveillance Act is set to lapse after Congress failed to pass either a long-term renewal or a temporary extension. The authority allows U.S. intelligence agencies to collect the communications of foreigners overseas without a warrant, and while existing court-approved collection is expected to continue temporarily, the government may be unable to obtain new authorisations.

⸻

🏛️ Government, procurement & public sector tech

France’s domestic intelligence service will replace Palantir’s AI data tools with those of French firm ChapsVision as part of a strategy to reduce dependence on foreign technology providers, though the transition is expected to take several years as Palantir’s contract was renewed in 2025 and runs until 2028. France also announced €655 million in AI investment and plans to expand government use of domestic AI systems.

Kazakhstan signed agreements with US startup Firebird Inc. on AI computing and infrastructure projects that could attract up to $10 billion in investment, including plans for a large-scale project known as Data Center Valley in northeastern Kazakhstan, with support involving Nvidia technologies.

In the United States, the Trump administration has positioned itself against formal AI regulation while increasingly shaping the sector through executive actions, export controls, procurement requirements and voluntary frameworks. With Congress stalled on AI legislation, the administration has become the primary driver of AI governance, and industry concerns are growing over policy uncertainty as companies navigate evolving government expectations without clear statutory rules.

⸻

🧒 Online harms & child safety

UK Prime Minister Keir Starmer announced a ban preventing under-16s from accessing major social media platforms including TikTok, Instagram and X, due to take effect in early 2027. The measures would also restrict features on other online services including chat functions with strangers, disappearing messages and location sharing, introduce limits on late-night scrolling for older teenagers, and ban under-18s from using romantic or sexual AI chatbots.

Major technology companies including Meta, YouTube and Snapchat criticised the policy, arguing it could drive young users towards less regulated platforms, while the UK government has begun lobbying US officials and technology companies to reduce the risk of retaliation over the planned restrictions.

Florida sued TikTok, alleging the platform violates state social media regulations for minors and consumer protection laws by understating the amount of adult content accessible to teenagers. TikTok says it has extensive safety features for teenagers and is updating its platform to comply with Florida law.

Roblox is expanding its age-verification system using AI-based facial age estimation, behavioural analysis and identity verification to improve child safety on the platform, though the measures have generated concerns about privacy, accuracy and barriers for users and developers.

Bloomberg examined the online network known as 764, a loose collection of cyberbullying and exploitation groups that authorities say coerce minors into self-harm, violence, sexual exploitation and criminal acts. The FBI has hundreds of active investigations linked to the network and has characterised it as a nihilistic violent extremist movement, with US law enforcement, Homeland Security and child-protection agencies pursuing members while working to disrupt associated Discord and Telegram communities.

⸻

🌏 Global policy

Australia

Shadow industry minister Andrew Hastie called for Australia to significantly increase investment in artificial intelligence, proposing the appointment of an AI ambassador, stronger AI education and positioning Australia as a southern hemisphere technology hub.

Chris Taylor, in his ASPI report and The Strategist article, argued that advances in generative AI are creating pressure on Australia’s National Intelligence Community to redesign intelligence products and delivery methods, with AI potentially enabling interactive querying of intelligence holdings, personalised outputs and faster dissemination, while also introducing risks related to accuracy, accountability and bias.

James Corera and Gatra Priyandita in their The Strategist article discussed Australia cybersecurity Horizon 2 Action Plan, released by the Department of Home Affairs on 11 June.

Submissions to a federal parliamentary inquiry describe growing online racism directed at Aboriginal and Torres Strait Islander people, with the Australian Human Rights Commission recommending a digital duty of care requiring platforms to assess and mitigate risks arising from recommender systems that amplify racist content.

A Vodafone network outage disrupted mobile services across Australia on 18 June, with owner TPG Telecom saying emergency Triple-0 access was maintained. The outage occurred amid ongoing regulatory scrutiny of Australian telecommunications networks following a September 2025 Optus outage that disrupted Triple Zero access, with new rules now requiring telcos to report outages to regulators and emergency services in real time.

Chile

Chile’s consideration of a proposed $500 million undersea cable linking Valparaíso and Hong Kong became entangled in US-China strategic competition after the US revoked visas of Chilean officials involved in assessing the project. Chile continues to weigh the proposal amid efforts to balance relations with China, its largest trading partner, and the United States.

Europe

Estonia plans to assign personal identification numbers to AI assistants to regulate the authority and access granted to AI systems acting on behalf of individuals and businesses — which would make it the first country to implement such a system for AI agents. The country will also introduce additional security screening for emails originating from Russia’s .ru domain before they reach government officials, automatically quarantining such messages as part of broader efforts to counter Russian hybrid threats.

Ukraine has been granted access to the European Union’s Cybersecurity Reserve, enabling it to request assistance from EU-approved cybersecurity experts during major cyber incidents. The reserve, managed by ENISA, integrates Ukraine into a collective European cyber defence mechanism despite its not yet being an EU member.

India

India temporarily blocked Telegram until 22 June, alleging the platform was used by organised groups seeking to defraud candidates sitting a re-examination for the National Eligibility cum Entrance Test, following the cancellation of results for 2.3 million students amid paper leak investigations. Authorities also ordered Google and Apple to remove Telegram from their app stores in India during the restriction period. Telegram subsequently challenged the block in court, arguing the measure is unlawful and disproportionate, and said targeted content removals would have been a less restrictive response.

United States

POLITICO found that 40 of 69 competitive US House districts have data centres planned or under construction, making AI infrastructure a growing political issue ahead of the midterm elections. Candidates from both parties are balancing economic and national security arguments for data centres against voter concerns over electricity prices, water use, land use and environmental impacts.

A Public First survey found growing pessimism about AI among Americans between 2024 and 2026, with more respondents now believing the technology will harm society than improve it. The survey also found that many respondents viewed China as better positioned than the US in the global AI race, while confidence in AI’s societal benefits was stronger in countries such as Singapore and India.

US security agencies are preparing for the 2026 FIFA World Cup across 11 host cities, with officials identifying lone-actor attacks, weaponised drones and AI-enabled disinformation as key security concerns, and the FBI having trained 60 officers in counter-drone tactics.

Companies are increasingly introducing policies governing employee use of prediction markets as concerns grow over insider trading and misuse of confidential information, with the rapid growth of platforms such as Kalshi and Polymarket expanding wagering into corporate events, earnings calls and product launches.

United Kingdom

Britain’s Competition and Markets Authority ordered Google to increase transparency in how its search rankings operate and to provide clearer complaint mechanisms for businesses, with the measures requiring objective criteria for organic search rankings and allowing users to transfer search data to authorised third parties. The regulator said the rules are intended to address Google’s dominance in UK search, where it handles more than 90% of queries.

---

That’s all for this week. For more timely analysis and commentary, check out The Strategist and ASPI’s Stop the World podcast—or our other Substack newsletters:

---

The Cyber & Tech Digest is brought to you by the Cyber, Technology & Security Programs team at ASPI and supported by partners.