cetas.turing.ac.uk/ Research Report
As AI increasingly shapes the global economic and security landscape, China’s ambitions for global AI dominance are coming into focus. This CETaS Research Report, co-authored with Adarga and the International Institute for Strategic Studies, explores the mechanisms through which China is strengthening its domestic AI ecosystem and influencing international AI policy discourse. The state, industry and academia all play a part in the process, with China’s various regulatory interventions and AI security research trajectories linked to government priorities. The country’s AI security governance is iterative and is rapidly evolving: it has moved from having almost no AI-specific regulations to developing a layered framework of laws, guidelines and standards in just five years. In this context, the report synthesises open-source research and millions of English- and Chinese-language data points to understand China’s strategic position in global AI competition and its approach to AI security.

This CETaS Research Report, co-authored with the International Institute for Strategic Studies (IISS) and Adarga, examines China’s evolving AI ecosystem. It seeks to understand how interactions between the state, the private sector and academia are shaping the country’s strategic position in global AI competition and its approach to AI security. The report is a synthesis of open-source research conducted by IISS and Adarga, leveraging millions of English- and Chinese-language data points.

Key Judgements
China’s political leadership views AI as one of several technologies that will enable the country to achieve global strategic dominance. This aligns closely with President Xi’s long-term strategy of leveraging technological revolutions to establish geopolitical strength. China has pursued AI leadership through a blend of state intervention and robust private-sector innovation. This nuanced approach challenges narratives of total government control, demonstrating significant autonomy and flexibility within China’s AI ecosystem. Notably, the development and launch of the DeepSeek-R1 model underscored China's ability to overcome significant economic barriers and technological restrictions, and almost certainly caught China’s political leadership by surprise – along with Western chip companies.

While the Chinese government retains ultimate control of the most strategically significant AI policy decisions, it is an oversimplification to describe this model as entirely centrally controlled. Regional authorities also play significant roles, leading to a decentralised landscape featuring multiple hubs and intense private sector competition, which gives rise to new competitors such as DeepSeek. In the coming years, the Chinese government will almost certainly increase its influence over AI development through closer collaboration with industry and academia. This will include shaping regulation, developing technical standards and providing preferential access to funding and resources.

China's AI regulatory model has evolved incrementally, but evidence suggests the country is moving towards more coherent AI legislation. AI governance responsibilities in China remain dispersed across multiple organisations. However, since February 2025, the China AI Safety and Development Association (CnAISDA) has become what China describes as its counterpart to the AI Security Institute. This organisation consolidates several existing institutions but does not appear to carry out independent AI testing and evaluation.

The Chinese government has integrated wider political and social priorities into AI governance frameworks, emphasising what it describes as “controllable AI” – a concept interpreted uniquely within the Chinese context. These broader priorities directly shape China’s technical and regulatory approaches to AI security. Compared to international competitors, China’s AI security policy places particular emphasis on the early stages of AI model development through stringent controls on pre-training data and onerous registration requirements. Close data sharing between the Chinese government and domestic AI champions, such as Alibaba’s City Brain, facilitates rapid innovation but would almost certainly encounter privacy and surveillance concerns if attempted elsewhere.

The geographical distribution of China's AI ecosystem reveals the strategic clustering of resources, talent and institutions. Cities such as Beijing, Hangzhou and Shenzhen have developed unique ecosystems that attract significant investments and foster innovation through supportive local policies, including subsidies, incentives and strategic infrastructure development. This regional specialisation emerged from long-standing Chinese industrial policy rather than short-term incentives.

China has achieved significant improvements in domestic AI education. It is further strengthening its domestic AI talent pool as top-tier AI researchers increasingly choose to remain in or return to China, due to increasingly attractive career opportunities within China and escalating geopolitical tensions between China and the US. Chinese institutions have significantly expanded domestic talent pools, particularly through highly selective undergraduate and postgraduate programmes. These efforts have substantially reduced dependence on international expertise, although many key executives and researchers continue to benefit from an international education.

Senior scientists hold considerable influence over China’s AI policymaking process, frequently serving on government advisory panels. This stands in contrast to the US, where corporate tech executives tend to have greater influence over AI policy decisions.

Government support provides substantial benefits to China-based tech companies. China’s government actively steers AI development, while the US lets the private sector lead (with the government in a supporting role) and the EU emphasises regulating outcomes and funding research for the public good. This means that China’s AI ventures often have easier access to capital and support for riskier projects, while a tightly controlled information environment mitigates against reputational risk.

US export controls have had a limited impact on China’s AI development. Although export controls have achieved some intended effects, they have also inadvertently stimulated innovation within certain sectors, forcing companies to do more with less and resulting in more efficient models that may even outperform their Western counterparts. Chinese AI companies such as SenseTime and DeepSeek continue to thrive despite their limited access to advanced US semiconductors.

• Initial access was via a password spray attack against an exposed RDP server, targeting numerous accounts over a four-hour period.
• Mimikatz and Nirsoft were used to harvest credentials, with evidence of LSASS memory access.
• Discovery was accomplished using living-off-the-land binaries as well as Advanced IP Scanner and NetScan.
• Rclone was used to exfiltrate data to a remote server using SFTP.
• The threat actor deployed RansomHub ransomware network wide, which spread over SMB and was executed using remote services.

Following major public exposures by Insikt Group and others throughout the last two years, alongside US government sanctions targeting the Intellexa Consortium — the organizational structure behind the Predator mobile spyware — Insikt Group observed a significant decline in Predator-related activity. This apparent decline raised questions about whether the combination of US sanctions, public exposure, and broader international efforts to curb spyware proliferation, such as the UK and France-led Pall Mall process, had dealt a lasting blow to Intellexa’s operations. Yet, Predator activity has not stopped, and in recent months, Insikt Group has observed a resurgence of activity, reflecting the operators’ continued persistence. While much of the identified infrastructure is tied to known Predator operators in countries previously identified by Insikt Group, a new customer has also been identified in Mozambique — a country not previously publicly linked to the spyware. This aligns with the broader observation that Predator is highly active in Africa, with over half of its identified customers located on the continent. Additionally, Insikt Group has found a connection between high-tier Predator infrastructure and a Czech entity previously associated with the Intellexa Consortium.

• Insikt Group has identified new infrastructure associated with Predator, indicating continued operations despite public exposure, international sanctions, and policy interventions.
• The newly identified infrastructure includes both victim-facing Tier 1 servers as well as high-tier components that likely link back to Predator operators in various countries.
• Although much of Predator’s infrastructure remains consistent with previous reporting, its operators have introduced changes designed to further evade detection — a pattern Insikt Group noted in earlier reporting.
• Insikt Group has detected Predator-related activity in several countries throughout the last twelve months and is the first to report a suspected Predator operator presence in Mozambique.
• Insikt Group also connected components of Predator’s infrastructure to a Czech entity previously linked with the Intellexa Consortium by a Czech investigative outlet.

The number of reported cyber incidents and online threats in Switzerland rose sharply last year, according to the National Cyber Security Centre (NCSC).

Last year, almost 63,000 cyber-related incidents were reported to the National Cyber Security Centre (NCSC) in Switzerland, an increase of 13,500 cases over the previous year. Between July and December, the NCSC recorded more than 28,000 incidents, slightly fewer than in the first half of 2024.

Fraud, phishing and spam messages continue to be the most frequently reported incidents. The increase on the previous year is mainly due to the phenomenon of false calls in the name of the authorities, with almost 22,000 reports compared with around 7,000 the previous year.

On the other hand, the number of e-mail threats has dropped. Over the past four years, fraudsters have used the telephone more as a communication channel.

This report highlights a rarely-discussed but crucially important attack surface: security vendors themselves.

• In recent months, SentinelOne has observed and defended against a spectrum of attacks from financially motivated crimeware to tailored campaigns by advanced nation-state actors.
• These incidents were real intrusion attempts against a U.S.-based cybersecurity company by adversaries, but incidents such as these are neither new nor unique to SentinelOne.
• Recent adversaries have included:
  • DPRK IT workers posing as job applicants
    ransomware operators probing for ways to access/abuse our platform
  • Chinese state-sponsored actors targeting organizations aligned with our business and customer base
    This report highlights a rarely-discussed but crucially important attack surface: security vendors themselves.

Malware Analysis Report - LockBit Ransomware v4.0

In this blog post, I’m going over my analysis for the latest variant of LockBit ransomware - version 4.0. Throughout this blog, I’ll walk through all the malicious functionalities discovered, complete with explanations and IDA screenshots to show my reverse engineering process step by step. This new version of LockBit 4.0 implements a hybrid-cryptography approach, combining Curve25519 with XChaCha20 for its file encryption scheme.

This version shares similarities with the older LockBit Green variant that is derived from Conti ransomware. While the multi-threading architecture seems more streamlined than previous versions, it still delivers an encryption speed that outpaces most other ransomware families.

As always, LockBit is still my most favorite malware to look at, and I certainly enjoyed doing a deep dive to understand how this version works.

This Google Threat Intelligence Group report presents an analysis of detected 2024 zero-day exploits.

Google Threat Intelligence Group (GTIG) tracked 75 zero-day vulnerabilities exploited in the wild in 2024, a decrease from the number we identified in 2023 (98 vulnerabilities), but still an increase from 2022 (63 vulnerabilities). We divided the reviewed vulnerabilities into two main categories: end-user platforms and products (e.g., mobile devices, operating systems, and browsers) and enterprise-focused technologies, such as security software and appliances.

Vendors continue to drive improvements that make some zero-day exploitation harder, demonstrated by both dwindling numbers across multiple categories and reduced observed attacks against previously popular targets. At the same time, commercial surveillance vendors (CSVs) appear to be increasing their operational security practices, potentially leading to decreased attribution and detection.

We see zero-day exploitation targeting a greater number and wider variety of enterprise-specific technologies, although these technologies still remain a smaller proportion of overall exploitation when compared to end-user technologies. While the historic focus on the exploitation of popular end-user technologies and their users continues, the shift toward increased targeting of enterprise-focused products will require a wider and more diverse set of vendors to increase proactive security measures in order to reduce future zero-day exploitation attempts.

April 23, 2025

The Federal Bureau of Investigation’s Internet Crime Complaint Center (IC3) has released its latest annual report. The 2024 Internet Crime Report combines information from 859,532 complaints of suspected internet crime and details reported losses exceeding $16 billion—a 33% increase in losses from 2023.

The top three cyber crimes, by number of complaints reported by victims in 2024, were phishing/spoofing, extortion, and personal data breaches. Victims of investment fraud, specifically those involving cryptocurrency, reported the most losses—totaling over $6.5 billion.

According to the 2024 report, the most complaints were received from California, Texas, and Florida. As a group, people over the age of 60 suffered the most losses at nearly $5 billion and submitted the greatest number of complaints.

“Reporting is one of the first and most important steps in fighting crime so law enforcement can use this information to combat a variety of frauds and scams,” said FBI Director, Kash Patel. “The IC3, which is celebrating its 25th anniversary this year, is only as successful as the reports it receives; that’s why it’s imperative that the public immediately report suspected cyber-enabled criminal activity to the FBI.”

To promote public awareness, the IC3 produces an annual report to aggregate and highlight the data provided by the general public. The quality of the data is a direct reflection of the information the public provides through the IC3 website. The IC3 standardizes the data by categorizing each complaint and analyzes the data to identify and forecast trends in internet crime. The annual report helps the FBI develop effective relationships with industry partners and share information for investigative and intelligence purposes for law enforcement and public awareness.

The IC3, which was established in May 2000, houses nine million complaints from the public in its database and continues to encourage anyone who thinks they’ve been the victim of a cyber-enabled crime, regardless of dollar loss, to file a complaint through the IC3 website. The more comprehensive complaints the FBI receives, the more effective it will be in helping law enforcement gain a more accurate picture of the extent and nature of internet-facilitated crimes.

The FBI recommends that everyone frequently review consumer and industry alerts published by the IC3. If you or your business are a victim of an internet crime, immediately notify all financial institutions involved in the relevant transactions, submit a complaint to www.ic3.gov, contact your nearest FBI field office, and contact local law enforcement.

Learn more about the history of IC3 by listening to this previously released FBI podcast episode: Inside the FBI: IC3 Turns 20.

This report details a newly identified and active fraud campaign, highlighting the emergence of sophisticated mobile malware leveraging innovative techniques:

• SuperCard X Malware: A novel Android malware offered through a Malware-as-a-Service (MaaS) model, enabling NFC relay attacks for fraudulent cash-outs.
• Evolving Threat Landscape: Demonstrates the continuous advancement of mobile malware in the financial sector, with NFC relay representing a significant new capability.
• Combined Attack Vectors: Employs a multi-stage approach combining social engineering (via smishing and phone calls), malicious application installation, and NFC data interception for highly effective fraud.
• Low Detection Rate: SuperCard X currently exhibits a low detection rate among antivirus solutions due to its focused functionality and minimalistic permission model.‍
• Broad Target Scope: The fraud scheme targets customers of banking institutions and card issuers, aiming to compromise payment card data.

The dawn of our fifth year deepens our understanding of the enemies at the gate, and some tensions inside it; plus, an anniversary gift from us to you

Inexpensive information-stealing malware surged in 2024, infecting 23 million hosts, according to Flashpoint.

Flashpoint data points to a surge in data breaches fueled by compromised credentials, ransomware and exploits

In a first-of-its-kind report, the US government has revealed that it disclosed 39 zero-day software vulnerabilities to vendors or the public in 2023 for the purpose of getting the vulnerabilities patched or mitigated, as opposed to retaining them to use in hacking operations.

It’s the first time the government has revealed specific numbers about its controversial Vulnerabilities Equities Process (VEP) — the process it uses to adjudicate decisions about whether zero-day vulnerabilities it discovers should be kept secret so law enforcement, intelligence agencies, and the military can exploit them in hacking operations or be disclosed to vendors to fix them. Zero-day vulnerabilities are security holes in software that are unknown to the software maker and are therefore unpatched at the time of discovery, making systems that use the software at risk of being hacked by anyone who discovers the flaw.

Since the end of 2024, we have been continuously monitoring large-scale DDoS attacks orchestrated by an IoT botnet exploiting vulnerable IoT devices such as wireless routers and IP cameras.

In 2024, ransomware groups claimed responsibility for 5,461 successful ransomware attacks on organizations worldwide. 1,204 of these attacks were confirmed by the targeted organizations. The rest were claimed by ransomware groups on their data leak sites, but have not been acknowledged by the targets.

A deep dive into macOS malware this year.

In this new report, learn how threat actors are leveraging cloud services to target web services with ransomware attackers.

A report by Secureworks revealed a 30% year-over-year rise in active ransomware groups, which demonstrates fragmentation of an established criminal ecosystem.

Version 1.1: October 18, 2024

• Based on our investigations, we are confident that there has been no breach of our systems.
• We have determined that the data in question is on a public-facing * DevHub environment—a Cisco resource center that enables us to support our community by making available software code, scripts, etc. for customers to use as needed.
• At this stage in our investigation, we have determined that a small number of files that were not authorized for public download may have been published.
• As of now, we have not observed any confidential information such as sensitive PII or financial data to be included but continue to investigate to confirm.
• Out of an abundance of caution, we have disabled public access to the site while we continue the investigation.
• Meanwhile, Cisco will engage directly with customers if we determine they have been impacted by this event.

A stack-based buffer overflow vulnerability exists in the OpenPLC Runtime EtherNet/IP parser functionality of OpenPLC _v3 b4702061dc14d1024856f71b4543298d77007b88. A specially crafted EtherNet/IP request can lead to remote code execution. An attacker can send a series of EtherNet/IP requests to trigger this vulnerability.