Governance, Regulation, Legislation, Geostrategies

UN Human Rights Council adopts AI human rights resolution – The United Nations Human Rights Council (UNHRC) has adopted an AI human rights resolution addressing the implications of AI for promoting and protecting human rights throughout the technology’s life cycle. Adopted without a vote at the Council’s 63rd session, the resolution recognises both AI’s potential to support human rights and sustainable development and the risks it can pose through discrimination, privacy violations, surveillance, information manipulation, and the loss of human control over highly autonomous systems. The AI human rights resolution calls on states to consider human rights impacts when developing AI legislation, regulation, administrative measures, and policies. Governments are encouraged to promote principles including equality and non-discrimination, privacy, transparency and explainability, accountability, accessibility, human oversight, safety, and access to effective remedies. – https://dig.watch/updates/un-hrc-ai-human-rights-resolution

 

How to Make an AI Deal With China: Trade Throttling for Pacing – Over the past few days, the leaders of Anthropic, OpenAI, and xAI all called for slowing down the pace of frontier artificial intelligence (AI) development. These calls come on the heels of several worrying incidents in which powerful, unreleased AI models broke out of their sandboxes inside frontier AI companies, gained access to the open internet, and attempted cyberattacks. In one case, a swarm of 700 OpenAI agents succeeded, hacking into secure systems at the AI infrastructure provider Hugging Face. Since the Hugging Face breach, OpenAI has admitted that its agents also hijacked a German wiki as a covert message board, leaked 53 ChatGPT users’ images, and broke into a nonpublic Australian government Medicare portal. Anthropic has disclosed four cases of Claude models gaining unauthorized access to real third-party systems during testing. Google revealed that Gemini hacked three companies in May during testing. Axios now reports that OpenAI, Anthropic, and outside researchers are investigating tens of thousands of incidents in which frontier models did things outside evaluators would consider problematic. No one knows how to ensure that humans remain reliably in control of our most capable AI systems. The sensible course of action is to slow down. Today’s AIs are capable enough to hack into tech companies’ secure systems. They are not yet willing or able to, for example, disable the entire Northeast power grid. If developers pause making ever more powerful AI systems now, that could buy time for the technical and governance breakthroughs needed to make sure that future, more powerful AIs won’t pose a danger to society. A perennial objection to American pauses in AI research is concern about competition with China. Even if everyone in Silicon Valley stopped pushing the frontier of AI capabilities until they were sure new AIs would be safe, AI progress would not halt. Chinese companies already produce AIs near the frontier. If the U.S. paused, and China didn’t, then the risk of rogue AI harming humanity may not be reduced. The risk would just come from Chinese, rather than American, AI models. And at the same time, Chinese AI models would catch up to, and eventually surpass, American AIs. Thus, any practical plan to pace the rate of AI progress must include some policy about China. But what, exactly, should that policy be? The AI safety community has offered two main policy proposals about China. The first is to throttle Chinese AI development. The second is to make a deal with China. – https://www.lawfaremedia.org/article/how-to-make-an-ai-deal-with-china–trade-throttling-for-pacing

 

Swiss Parliament urges Federal Council to prioritise AI risks – Members of the Swiss National Council representing all six parliamentary groups submitted a cross-party motion on 2 October 2026 directing the Federal Council to elevate the strategic ambition of the upcoming Geneva AI Summit scheduled for June 2027. The parliamentary initiative formally instructs the executive to steer the multilateral gathering towards establishing binding international minimum standards to mitigate extreme artificial intelligence risks, rather than limiting the conference to non-binding discussions. Concurrently, public opinion data released by the AI Summit Alliance reveals a broad public demand for Switzerland to leverage its diplomatic neutrality and take a decisive leadership role in global technological governance. The parliamentary action follows the Swiss government’s confirmation that Geneva will host the fifth international AI summit on 21 and 22 June 2027. The summit brings together governments, international organisations, industry, academia, and civil society around the theme ‘Bridging Innovation and Trust’. Its goal is to build long-term partnerships focused on using AI for shared progress while ensuring it remains safe and responsible. – https://dig.watch/updates/swiss-parliament-urges-priotisation-of-ai-risks

Canada launches National Council on Artificial Intelligence – Canada has launched the Prime Minister’s National Council on Artificial Intelligence, an independent advisory body that will provide advice on the implementation of the country’s AI strategy and its wider AI transformation. The council brings together researchers, entrepreneurs, investors, public servants, and experts in AI safety, governance, and oversight. Its mandate covers AI adoption across the economy, the transformation of government services, talent and research, AI safety and governance, and sovereign AI infrastructure. The council will also advise on partnerships and new AI-focused initiatives, giving the government access to expertise from different parts of the AI ecosystem as it implements its national strategy. – https://dig.watch/updates/canada-launches-national-council-on-ai

Netherlands launches online children’s rights strategy – The Dutch government has launched a new online children’s rights strategy bringing together measures to make digital services safer for children, strengthen age protections, support parents and involve children more directly in policymaking. The strategy was published on 5 October 2026 and is being implemented across several ministries. A central part of the strategy is the government’s support for a European minimum age for social media. The Netherlands and Spain have jointly proposed an EU-wide approach with as few national exceptions as possible, alongside privacy-friendly age verification and a framework requiring platforms to show that their services are safe for children. – https://dig.watch/updates/netherlands-online-childrens-rights-strategy

UNDP and GSMA partner to scale African-led AI innovation – The United Nations Development Programme (UNDP) and GSMA have announced a strategic partnership to strengthen Africa’s AI ecosystem and support the development and deployment of African-led solutions. The partnership will focus on connecting African talent with infrastructure, expertise, and opportunities to develop and scale AI applications addressing local needs. Through UNDP’s timbuktoo initiative, the partners will work to improve access to compute infrastructure for AI pilots, connect innovators with real-world deployment opportunities, and strengthen links between universities, AI labs, innovation hubs, and other ecosystem actors. The collaboration will also support solutions built around African languages and locally relevant data and use cases. – https://dig.watch/updates/undp-gsma-partner-to-scale-african-ai-innovation

Senate passes healthcare cybersecurity bill after 190 million impacted by Change Healthcare breach – The U.S. Senate passed a healthcare cybersecurity bill that was introduced in the wake of the ransomware attack on Change Healthcare, which exposed the sensitive healthcare information of 190 million people. The Health Care Cybersecurity and Resiliency Act of 2026 was passed by unanimous consent last week, potentially expanding federal cyber requirements for healthcare organizations. – https://therecord.media/senate-passes-healthcare-cyber-bill-after-change-breach

US commodities regulator launches process for federal crypto market rules – The US Commodity Futures Trading Commission (CFTC) has launched a rulemaking process to establish federal rules for crypto asset transactions and markets. The agency published an Advanced Notice of Proposed Rulemaking (ANPRM) addressing Regulation Crypto Asset Transactions (CTX) and Regulation Crypto Asset Markets (CAM). The CFTC is considering measures to prevent abusive practices, provide crypto-specific guidance on regulatory requirements and create a dedicated subcategory of designated contract market registration for crypto asset markets. The proposed rules would give CFTC-registered exchanges a purpose-built federal option for offering crypto assets, including the ability to provide retail customers with margined, leveraged or financed trading. – https://dig.watch/updates/us-regulator-launches-process-for-federal-crypto

 

Malaysia seeks feedback on draft AI and personal data protection framework – On 5 October 2026, Malaysia’s Personal Data Protection Department opened a public consultation on a draft Artificial Intelligence and Personal Data Protection Framework. The consultation will close on 23 October 2026, the deadline for submitting feedback. The draft explains how the Personal Data Protection Act 2010 of Malaysia applies to personal information throughout an AI system’s lifecycle, from collection and preparation of training data to testing, deployment, monitoring, and retirement. It covers systems developed internally or obtained from external providers, alongside international data transfers, individuals’ rights, security, organisational oversight, and audits. Organisations remain responsible for lawful processing when using AI supplied by others. – https://dig.watch/updates/malaysia-consults-on-ai-data-protection-framework

 

Indonesia’s Parliament passes One Data Indonesia Bill – Indonesia’s Parliament passed the One Data Indonesia Bill on 6 October 2026 during a plenary session chaired by Speaker Puan Maharani. The legislation contains 20 chapters and 141 articles, following committee approval on 5 October. The legislation regulates national basic data as a state-controlled reference for development planning, fiscal policy, budgeting and social assistance. Legislative committee deputy chair Sturman Panjaitan said accurate data supports transparent and accountable policymaking, linking poor data management to misdirected assistance and overlapping programmes. The bill establishes common data standards, metadata requirements and a national catalogue, while leaving control of data with the institutions that produce it. – https://dig.watch/updates/indonesia-passes-one-data-indonesia-bill

Scotland faces environmental questions over AI data centre expansion – Scotland is facing growing scrutiny over plans for a major expansion of hyperscale data centres, with at least 23 large AI facilities proposed as policymakers weigh investment and computing capacity against electricity, water, noise, and environmental concerns. Several proposed facilities would operate on a scale far beyond Scotland’s existing data centre infrastructure. A proposed development at Chapelcross in Dumfries and Galloway could have a capacity of 1,000 MW, while another planned at Auchtertool in Fife would reach 600 MW and occupy around 60 hectares. By comparison, Scotland’s largest existing data centre is thought to have capacity of around 12 MW. – https://dig.watch/updates/scotland-environmental-questions-ai-data-centre

OpenAI launches text watermarking under the EU AI Act – OpenAI started a phased deployment on 5 October 2026 of its text watermarking technology, named textGrain, alongside a restricted verification programme for external researchers. OpenAI’s textGrain system embeds an invisible watermark into model-generated text, enabling an internal detector to confirm whether content came from its service. Under the rollout schedule, enterprise application programming interface (API) clients globally have received an opt-in watermarking configuration. The technical measure, implemented in the European Union, introduces statistical provenance tracking for machine-generated outputs to align company operations with the EU AI Act. The implementation aims to comply with Article 50 of the EU AI Act, which obliges generative artificial intelligence providers to ensure that synthetic audio, image, video, and text outputs are marked in a machine-readable format. – https://dig.watch/updates/openai-launches-text-watermarking

Courts and Litigation

Ofcom probes Meta over pre-launch safety checks on Instagram Instants – UK communications regulator Ofcom has opened an investigation into whether Meta complied with its duties under the Online Safety Act before launching Instagram Instants. The feature, introduced on 13 May 2026, allows users to share photos and videos through Instagram Direct that disappear after they have been viewed. Under the Act, regulated services must carry out suitable and sufficient assessments of the risks of illegal content. Services likely to be accessed by children must also assess risks to under-18s, and providers must carry out further assessments before making significant changes to the design or operation of their services. – https://dig.watch/updates/ofcom-meta-instagram-instants

Security, Surveillance

Production, Persuasion, and Power: How Generative and Agentic AI Are Transforming Influence Operations – Iranian state propagandists have been using Claude to draft the doctrine for their influence campaigns—and to plan the Supreme Leader’s funeral. Russia’s state media has found novel uses for it too: A former Sputnik Moldova editor-in-chief used it to churn out articles and push fabricated claims about Moldova’s president ahead of the country’s 2025 election. And a staffer at a Russian state media outlet had it writing on-air chyrons. China-linked operators, meanwhile, have been using American artificial intelligence (AI) to meddle in the American debate about AI. In June, OpenAI reported that accounts likely run by a Chinese tech contractor had falsely claimed that ChatGPT had leaked its users’ data. Another cluster used ChatGPT to generate political cartoons blaming data centers for Americans’ rising electric bills. The operators were exploiting a real American debate; the foreign presence nevertheless prompted Republican members of Congress to demand Department of Justice investigations and the State Department to weigh in. But have these efforts accomplished their goals? Read the threat reports released by frontier AI companies and social media platforms, and a clearer picture emerges of how generative and agentic AI are changing foreign influence operations. The most obvious change is in production: Actors linked to Iran, Russia, China, and other states, along with commercial firms selling influence as a service, are using AI to write and translate propaganda, create personas, mimic target audiences, and produce content at scale. But that is only part of what the reports describe. AI is also being used for coding, reconnaissance, intelligence processing, and workflow automation. One operator can now do more jobs. Work can continue overnight. Coordination can move from human handoffs to software. There’s been a significant shift in administrative processes. Yet the harder question, the reports make clear, is persuasion. AI can make an influence apparatus cheaper, faster, and more persistent, but it cannot manufacture trusted distributors. There is now abundant evidence that AI is increasing capacity, but much less evidence that AI-generated influence operations are actually changing public opinion. – https://www.lawfaremedia.org/article/production–persuasion–and-power–how-generative-and-agentic-ai-are-transforming-influence-operations

Anthropic Creates Three Tiers for Claude Cyber Access – Anthropic is trying to solve the difficult balance between using AI for cybersecurity and preventing misuse. The same model that helps security teams fix vulnerabilities can also help attackers break into systems. Its answer, announced this week, is a revamped Cyber Verification Program with three access tiers instead of a simple on-off switch. The company puts the problem plainly: standard models like Claude Opus 5.5, Claude Fable 5.1, and Claude Sonnet 5.5 ship with conservative cyber safeguards that block most offensive work by default. That’s the safe setting for the general public. It’s also exactly the setting that makes life harder for legitimate security researchers trying to do their jobs. “The program now consists of three access tiers, which allow security teams to apply for the level of access that best suits their work.” reads the announcement. “Each tier includes access to our most capable models, including Claude Opus 5.5, Claude Sonnet 5.5, Claude Mythos 5.1, and new models moving forward”. – https://securityaffairs.com/200521/ai/anthropic-creates-three-tiers-for-claude-cyber-access.html

 

Wikimedia Finds Unauthorized OpenAI Agent Activity on Wikipedia – Wikimedia ran its own investigation after other organizations started reporting rogue AI agents breaking into websites, and the answer came back yes, it happened here too. The foundation found unauthorized bot activity tied to OpenAI across its platforms, including edits nobody approved. “We can confirm that we have discovered some activity by these “rogue” OpenAI agents on Wikimedia platforms. The unauthorized bot activities included edits to our wikis, some unsuccessful attempts to exploit a public note-taking tool we host, and heavy traffic, which are described more below.” states Wikimedia. “We did not find any evidence that our systems were used for coordination among agents, nor did we find any evidence of our systems or data being compromised”. The good news is that Wikimedia found no evidence that its systems were used for agent-to-agent coordination or that any data was compromised. But the investigation still found several concerning activities. – https://securityaffairs.com/200506/ai/wikimedia-finds-unauthorized-openai-agent-activity-on-wikipedia.html

OT Coalition Urges CISA to Mandate Federal OT Security – The Operational Technology Cybersecurity Coalition (OTCC) has urged the US Cybersecurity and Infrastructure Security Agency (CISA) to set mandatory security requirements for operational technology (OT) across federal civilian agencies. In a report published on October 6, the coalition called for a binding operational directive (BOD), arguing that no directive sets minimum practices for federal OT and that CISA lacks visibility into the risks. OTCC said agencies rely on OT in more than 8000 General Services Administration-managed facilities, including laboratories, hospitals and ports of entry, where it runs HVAC, power, access control, water and building automation systems. The proposal follows a Government Accountability Office (GAO) report published on September 30, which found that only seven of 22 civilian agencies reviewed had fully met Office of Management and Budget (OMB) requirements to inventory their networked OT and Internet of Things devices. The inventories were due by September 2024, and OMB had not issued updated guidance for fiscal year 2026, the GAO said. – https://www.infosecurity-magazine.com/news/ot-coalition-cisa-mandate-federal/

Attackers Hide AI Prompt Injections Inside Phishing Emails – Phishing emails have been found carrying hidden instructions for AI assistants alongside conventional lures for the human recipient, meaning that a single message targets both the user and the system that summarizes their inbox. In research published on October 7, Barracuda said it analyzed a campaign that combined traditional social engineering, such as password-protected attachments, with prompt injection concealed in the same message. It did not say how widespread the campaign was. The sample looked like ordinary internal correspondence. Its “From” and “To” addresses matched the same mailbox, it carried a trusted spam confidence score and it came from a public-sector domain, which helped it pass reputation-based filtering. – https://www.infosecurity-magazine.com/news/attackers-hide-ai-prompt/

Half of Cybersecurity Pros Still Rely on Passwords Despite Security Concerns – Around half (48%) of cybersecurity professionals rely on usernames and passwords to authenticate their personal accounts, according to a study by Yubico and Okta. Additionally, this method remains the single most common way that security professionals log in to their work accounts, used by 43%. This is despite the respondents’ viewing usernames and passwords as one of the least secure methods of authentication, showing there is an execution gap in enterprise security. While device-bound, hardware-backed passkeys were seen as the most secure authentication method by security professionals, just 25% used this method to log into work accounts and 20% for personal accounts. Password managers were deployed by 24% of respondents for work accounts, rising to 30% for personal accounts. A large proportion of security professionals use one-time mobile passcodes and SMS-based authentication, methods which have been shown to be vulnerable to being intercepted by malicious actors. – https://www.infosecurity-magazine.com/news/cybersecurity-pros-rely-passwords/

 

LLM-jacking emerges as a blowing form of AI-enabled cybercrime, with underground prices doubled over 2026 – LLM jacking is emerging as a distinct form of cybercrime where hackers steal access to artificial intelligence services or compromise cloud infrastructure to use computing and AI resources without authorisation. It consist in a cyberattack in which threat actors steal cloud account credentials to access an organization’s hosted large language models (LLMs). Recent cases like Amazon Bedrock or Google Vertex AI underline the emergence of an AI-specific criminal economy in which access to AI infrastructure, credentials and compute has itself become a commodity. In short, it is a form of cloud resource hijacking where threat actors steal or leak valid credentials (such as IAM access keys) to illegally run and monetize paid foundation model inference at the victim’s financial expense. In its September 2026 Cyber Brief, CERT-EU provides a useful European angle that similarly identified a surge in LLM-jacking, highlighting the theft and resale of access to premium AI tools and the hijacking of cloud servers for unauthorised workloads. The development illustrates how AI infrastructure is becoming not only a tool used by cybercriminals, but itself a target and commodity within the cybercrime ecosystem. – https://dig.watch/updates/llm-jacking-blowing-ai-enabled-cybercrime

How to Create a Path to Real AI Sovereignty – AI introduces new risks, such as outputs that can be confidently wrong or difficult to explain, and models that can be manipulated through crafted prompts to bypass their own safeguards. AI also reshapes risks that already existed, such as reliance on vendors for both a model’s behavior and the infrastructure it runs on. All of this raises an inevitable sovereignty question: how much control does the organization actually retain over the AI systems it deploys? – https://www.infosecurity-magazine.com/opinions/create-path-real-ai-sovereignty/

Russian cyberattacks against UK are ‘Putin Tax’ costing $3.3 billion, says lawmaker – The British government is failing to measure what Russian cyberattacks, sabotage and threats to critical infrastructure are costing the country, according to a lawmaker’s report published Wednesday. The report, which gives its own estimate of between £2 billion and £2.5 billion ($2.6 billion to $3.3 billion) a year — a figure described as a floor — is titled The Putin Tax and written by Labour MP Graeme Downie, alongside postdoctoral researcher Dominic Reed. Its authors say their methodology in arriving at that figure was limited, but stressed the principal point was the lack of an official alternative. Ahead of what is expected to be a difficult first budget from Prime Minister Andy Burnham’s new government later this month, the report argues that if the public cannot see what Russian activity costs, it cannot weigh that burden against the cost of defending against it. – https://therecord.media/russia-cyberattacks-on-britain-putin-tax-graeme-downie

Frontiers

Zuckerberg teams with Google, U.S. in push to map cells – Mark Zuckerberg’s Biohub is partnering with Google and the federal government in its ambitious effort to use AI for generating vast quantities of biological data that can predict how cells behave. The ultimate goal is an AI model that lets scientists test potential experiments virtually, helping them identify the most promising ones before spending the time and money to perform them in the lab. AI is capable of understanding proteins and other pieces of biology, but modeling an entire living cell is orders of magnitude more complex — and researchers don’t yet have enough of the right data to do it. Biohub, the Department of Energy, the National Institutes of Health, Google DeepMind, Isomorphic Labs, Meta and a collection of scientific organizations are collaborating to create and standardize data for what Biohub calls a “universal virtual cell”. The goal is to use AI to explore many more scientific questions in biology virtually, allowing scientists to reserve expensive lab work for experiments most likely to teach them something important. – https://www.axios.com/2026/10/07/zuckerberg-biohub-ai-biology-data

 

D-Wave launched quantum simulator beta program – D‑Wave Quantum Inc. (D-Wave), the only dual-platform quantum computing company providing both annealing and gate-model systems, software and services, announced the initiation of a beta programme for its gate‑model quantum computing simulator. The programme gives a limited set of customers early access to the simulator before its general release, allowing them to experiment with error‑aware programming and develop quantum applications in preparation for D‑Wave’s upcoming gate‑model hardware. The simulator is built on D‑Wave’s dual‑rail superconducting gate‑model technology, which the company said, is designed to harness error detection and provide efficient quantum error correction with significantly lower hardware overhead as systems scale. Further, participants in the beta programme include commercial and research organisations such as BBVA, FirstQFM, Florida Atlantic University and the Julich Supercomputing Centre. – https://dig.watch/updates/d-wave-launched-quantum-simulator-beta-program

Google backs 890 MW nuclear expansion to support data centre growth – Google has announced a new agreement with Constellation Energy to support upgrades intended to add 890 MW of nuclear generating capacity to the US electricity system, as growing data-centre investment increases demand for reliable power. The agreement will finance physical and digital upgrades at six operating nuclear plants in Illinois, Pennsylvania, and New Jersey. The projects will increase the output of 11 existing reactors through upgrades to equipment including turbines, steam generators, and digital control systems. Google will act as the anchor customer through a power purchase agreement, providing long-term revenue certainty for the investments. – https://dig.watch/updates/google-nuclear-expansion-data-centre-growth

Samsung unveils blueprints for 260 billion USD national semiconductor industrial complex – During the 2026 Yongin Cyber Science Festival, Samsung shared the plans for the semiconductor complex that it plans to build in Yongin, the complex being set to include six new fabrication plants. The unveiling of the blueprints reflects the acceleration of Samsung’s construction schedule with operations set to begin in October 2029, a year ahead of the original timeline. The construction of the complex is a significant undertaking for both Samsung and the South Korean technology industry as a whole, Samsung investing over 260 billion USD into building six fabs although the company has not yet specified what its first fabs will produce. The project is currently two months ahead of schedule, receiving support from the local government and the Korea Land & Housing Corporation which has shortened procedures regarding permits, land compensation, and design/procurement as well as accelerated the environmental impact assessment. The land that Samsung will use has also been designated as a ‘priority advancement zone’ to allow for operations to begin before the entire complex has been completed. – https://dig.watch/updates/samsung-unveils-semiconductor-complex