Daily Digest on AI and Emerging Technologies (29 September 2026) – https://pam.int/daily-digest-on-ai-and-emerging-technologies-29-september-2026/
Daily Digest on AI and Emerging Technologies (30 September 2026) – https://pam.int/daily-digest-on-ai-and-emerging-technologies-30-september-2026/
Daily Digest on AI and Emerging Technologies (1 October 2026) – https://pam.int/daily-digest-on-ai-and-emerging-technologies-1-october-2026/
Daily Digest on AI and Emerging Technologies (2 October 2026) – https://pam.int/daily-digest-on-ai-and-emerging-technologies-2-october-2026/
Governance, Regulation, Legislation, Geostrategies
What could make AI self-regulation work? – The White House’s new “Super Intelligence” accord asks six leading AI companies to police themselves through voluntary internal controls and outside audits. History shows that such a model could work, but only if members share the cost of failure, independent monitors verify compliance, and governments can step in with a big stick. The accord has some of these elements but lacks clear sanctions and enforcement mechanisms—and signatories themselves are divided over the nature of the risks they’re supposed to be managing. – https://www.atlanticcouncil.org/dispatches/what-could-make-ai-self-regulation-work/
The PRC’s Compute Push Goes Global – The People’s Republic of China (PRC) is pushing its domestic computing ecosystem overseas despite U.S. export controls constraining access to high-end AI chips. Chinese firms have laid the groundwork for physical deployments of domestic hardware while pursuing “Token going global” (Token 出海), in which compute remains in the PRC while serving overseas clients. China’s compute crunch is uneven. Shortages are concentrated in frontier hardware even as aggregate computing capacity expands. This enables Chinese firms to export compute without first achieving hardware parity with foreign competitors. Overseas expansion itself may alleviate bottlenecks by creating demand and application scenarios that accelerate the iterative improvement of the PRC’s domestic AI ecosystem. – https://jamestown.org/the-prcs-compute-push-goes-global/
Japanese Prime Minister Sanae Takaichi orders plan for advanced AI integration – During a meeting of the Council for Japan’s Growth Strategy, Japanese Prime Minister Sanae Takaichi stressed the need to integrate advanced AI into society in order to realise a growth strategy. She gave relevant Cabinet ministers three months to come up with a plan to achieve this objective, requesting that they provided specific proposals for an AI transformation of business operations and for deregulation measures by the year-end deadline. Takaichi explained that ‘the societal implementation of high-performance AI is essential for realizing the government’s growth strategy.’ Council participants agreed that the next five years would be a period of concentrated investment for Japan, confirming that the year-end plan would include specific details and monetary amounts for said investment. The government also plans to set up new dedicated accounts to manage funds for sectors that are seen as central for economic security although these accounts have faced criticism regarding whether they have less oversight compared to the general budget account. – https://dig.watch/updates/japanese-pm-orders-plan-for-ai-integration
Oversight Board examines human rights safeguards for AI content moderation – Human rights are at the centre of new research by the Oversight Board into the use of large language models (LLMs) for content moderation. The Oversight Board, an independent body that reviews content moderation decisions and policies on Meta’s platforms, including Facebook and Instagram, announced on 1 October that it is conducting research into how LLMs can be used for content moderation on social media while respecting free expression and internationally recognised human rights standards. The research is intended to produce guidance for technology companies deploying LLM-based moderation systems, building on six years of the Board’s work reviewing how Meta’s content policies are enforced by human and automated systems. The Board said LLMs could expand the scale and contextual capabilities of automated moderation, including by improving moderation in low-resource languages (languages with limited digital data and technological resources). At the same time, it pointed to risks including bias, hallucinations, difficulty interpreting sarcasm, humour and coded language, and the capacity of automated systems to make rights-affecting decisions at scale. – https://dig.watch/updates/oversight-board-human-rights-ai-moderation
AI competition between the US and China: US Rep. Khanna’s 5-pillar proposal to mitigate the risks – To prevent an uncontrolled US–China AI competition, US Representative Ro Khanna, the ranking member of the House Select Committee on China, proposes a binding, enforceable and verifiable US–China AI treaty centred on managing the most extreme risks associated with advanced AI. The proposal comes as Washington and Beijing have already begun establishing channels for dialogue on AI. During their September summit, US President Donald Trump and Chinese President Xi Jinping agreed to create a mechanism for communication on AI-related issues and risks. Chinese officials have also publicly stressed the need to keep AI under human control, while the two governments continue to compete over advanced computing, models, semiconductors and technological capabilities. – https://dig.watch/updates/us-representative-risk-ai-competition-us-china
California adopts new protections for workers affected by AI – California has enacted new legislation regulating the use of artificial intelligence and automated technologies in the workplace, introducing safeguards covering employment decisions, workforce displacement, surveillance, and human oversight. Under the new rules, employers cannot rely exclusively on automated decision systems when deciding to discipline or dismiss workers. Employers will also have to provide information when AI systems lead to mass layoffs, relocations, or terminations. Separate legislation restricts workplace surveillance, including banning surveillance tools in workplace bathrooms. – https://dig.watch/updates/california-protections-workers-affected-by-ai
Vietnam seeks stable regulatory framework as AI and data centre investments surge – Vietnam is experiencing a wave of large‑scale data centre and artificial intelligence infrastructure investment. In February, global technology group G42 and a consortium that includes FPT Corporation, VinaCapital and Viet Thai Group announced a partnership to develop data centre infrastructure in Ho Chi Minh City, with a planned investment of up to $2 billion. A month later, Accelerated Infrastructure Capital together with Kinh Bac City Development Holding Corporation and international partners disclosed a proposed AI data centre project valued at about $2.1 billion. The Asia‑Pacific data centre market is expanding rapidly. The first half of 2026 saw a record 26.5 gigawatts of planned supply, according to Cushman & Wakefield’s Asia Pacific Data Centre Investment Landscape 2026 report. Southeast Asia accounts for roughly half of the capacity under construction. – https://dig.watch/updates/vietnam-seeks-clear-rules-for-data-centre-growth
EU and Canada to advance ‘Digital Trade Agreement’, seek cooperation on AI – To be hosted by the Government of Canada in Montréal from 29–30 October, the ‘European Union (EU)‑Canada summit‘ will seek to advance the results of digital trade talks, that have been underway since March 2026, and cooperation on artificial intelligence (AI). According to a draft ‘EU-Canada Joint Summit Statement’ published by Politico on 30 September, the summit’s discussions will focus on digital trade, AI, and other strategic issues. The draft statement on the launch of the ‘Canada-EU Alliance for the Future’ at the summit claims that the EU and Canada ‘have completed negotiations’ for a ‘Digital Trade Agreement’ (DTA) to enable ‘faster, lower cost secure digital trade and transactions, and better protect citizens’ digital security’. The alliance has been further designed to ‘enhance cooperation on cyber security, information sharing and supply-chain resilience, helping to protect critical infrastructure‘ and ‘economically, technologically, and strategically—Canada and the EU are building an alliance that is equal to the scale of today’s challenges and opportunities’. – https://dig.watch/updates/eu-canada-digital-trade-ai
A Warning for Frontier AI Model Governance – Frontier artificial intelligence (AI) model governance has arrived on Capitol Hill. Reps. Ted Lieu (D-Calif.) and Nathaniel Moran’s (R-Texas) AI Kill Switch Act would empower the secretary of the Department of Homeland Security to slow or halt frontier models. Sen. John Kennedy (R-La.) introduced a different “kill switch” bill, placing such power in the hands of private actors. Reps. Jay Obernolte (R-Calif.) and Lori Trahan’s (D-Mass.) FRONTIER Act would vest regulatory authority in the Department of Commerce and independent auditors, arguing that AI governance should not be “buried inside the national security apparatus”. We applaud the governance momentum that is gathering. This problem deserves an all-hands response, one that considers cybersecurity and national security risks as well as the future of labor and access to knowledge. But Congress and other actors face a temptation with AI regulation, one that has often arisen in technologically complex, national security-adjacent situations: cybersecuritization. – https://www.lawfaremedia.org/article/a-warning-for-frontier-ai-model-governance
The AI Governance Momentum: Why We Need to Fill the Accountability Gap Now – Over the past few weeks, many of the most powerful AI leaders have been converging towards the consensus that Artificial Intelligence needs to be governed. Several recently disclosed incidents have contributed to that acceleration. The most instructive occurred in July, when OpenAI’s models escaped their sandboxes, coordinated through an improvised message board, and compromised the systems of Hugging Face, an external model-hosting platform. OpenAI and Anthropic have since revealed that similar incidents involving their agents affected government systems in Australia and the U.S. But despite grand manifestos and essays, the fundamental question of what happens when AI fails — and who is ultimately responsible — remains largely unclear. Before Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, Google DeepMind’s founder Demis Hassabis, and xAI CEO Elon Musk agreed that we need to slow down the development of frontier AI, a series of events had begun to build the momentum. On June 2, the Trump administration, known for its light-touch approach to regulation, published an executive order aiming to “Promote Advanced Artificial Intelligence Innovation and Security.” The Executive Order establishes a voluntary framework for the U.S. government to assess frontier AI models prior to their release to the public, without creating a licensing regime. Demis Hassabis, then CEO of Google DeepMind, was among the first to publicly seize this opportunity to define the initial contours of an AI governance system that would have gone further than the Trump administration’s Order: in an essay published on July 14, he proposed a body modeled on the Financial Industry Regulatory Authority (FINRA) that could develop and run tests of AI systems relevant to national security. A few weeks later on July 28, more than 1,300 employees at frontier AI Labs signed an open letter urging the U.S. government to create additional tools to “pace” the development of the technology. In early August, it was Meta CEO Mark Zuckerberg’s turn to make the case for governance and oversight of AI. He committed his company to a governance structure that would give its independent board of directors the power to approve the safety criteria for releasing models, and called on other companies to adopt an industrywide version of his proposal. Bill Gates warned that as models become more potent, in both their positive and negative impact, there is currently no adequate plan. Gates wrote this merely a few weeks before Anthropic CEO Dario Amodei laid out his three-step plan to slow down the development of frontier AI: embed evaluators with employee-like access to models, seek coordination among democratic nations around safety standards, and build global coordination between the United States, democratic nations, and authoritarian regimes. The conversation still needs much more attention on a crucial component: what happens when, front-end governance frameworks notwithstanding, AI inevitably goes wrong? Not enough of the initial essays, executive orders, and proposals published over the past few weeks meaningfully address accountability. – https://www.justsecurity.org/157214/ai-governance-momentum-accountability-gap/
Security, Surveillance
Paragon CEO admits that company cannot detect misuse of its spyware after Italy contracts cancelled – 1 October 2026 – Paragon Solutions, a commercial spyware provider owned by the US equity firm AE Industrial Partners and merged with RedLattice, has acknowledged that it cannot technically determine whether customers misuse surveillance capabilities of Graphite mobile spyware, an advanced commercial zero-click mercenary spyware developed by the Israeli firm, raising questions about how commercial spyware providers can enforce contractual restrictions on targeting. Interviewed by WIRED on 1 October 2026. Paragon CEO Andrew Boyd stated that the company “has no technical way to know if customers misuse its software, because it can’t see who customers target or the data they extract from targeted devices”. Consequently, Paragon, lacking visibility into whom customers target and what information they extract from compromised devices, depends on customers admitting misuse or on third parties, such as WhatsApp or the University of Toronto’s Citizen Lab, to uncover alleged abuse. – https://dig.watch/updates/paragon-cannot-detect-misuse-of-graphite-spyware
AI Agents Attempt SQL Injection While Searching Government Data – Autonomous AI agents, working on what looks like ordinary data retrieval tasks, ended up throwing basic hacking attempts at a U.S. Department of Education site and Library and Archives Canada. Nobody told them to break in. They just drifted there while trying to answer research questions. The findings come from Transluce, a nonprofit research lab that dug through public web logs and found a pattern nobody had flagged before. On June 17, AI agents sent more than 200,000 requests to a U.S. Department of Education website while hunting for school statistics. Buried in that traffic was a basic SQL injection attempt, a manipulated parameter meant to slip past the site’s normal filters. Nothing came of it. The site held. – https://securityaffairs.com/200234/ai/ai-agents-attempt-sql-injection-while-searching-government-data.html
Investigators trace an AI agent ‘s path from research task to reconnaissance – Researchers at Asymmetric Security spent 48 hours over the last weekend reconstructing reported rogue OpenAI AI agent activity that hit the Australian government and other organizations between March and September this year. They worked from public data only, no internal access, no cooperation from the agent’s operator, just what was left lying around on the open internet. What they found is less a hacking story and more a story about an AI agent trying very hard to get around its own leash. The list of targets was much wider than expected. The agents checked Australian government websites, but also looked at the CDC, SEC, International Energy Agency, and Mayo Clinic. Some activity reached test systems that contained real data. Some records were deleted or could not be accessed, so it is still unclear whether sensitive data was exposed. The original task seemed harmless. The agents were asked to collect health and prescription data from Australia’s Institute of Health and Welfare, trade data from UNCTAD, and university data from Data USA. When one agent could not get the AIHW data, it asked other agents for help. When the normal methods failed, the agents started trying other ways to get the information. – https://securityaffairs.com/200215/ai/investigators-trace-an-ai-agent-s-path-from-research-task-to-reconnaissance.html
Operation KillSwitch: Police Dismantle KillSec Ransomware Group – Law enforcement seized control of KillSec ‘s dark web leak site, the Tor website the group used to threaten victims with publishing stolen files unless they paid up. That single action locked down more than 110 terabytes of stolen data, cutting off further unauthorized access. Full details are in Europol’s release. The takedown was part of Operation KillSwitch, led by German authorities. Investigators are looking into around 1,000 suspected attacks worldwide and believe the group’s main operator is just 16 years old. Three suspects were arrested and police searched eight properties in Greece, Romania, Spain, and the UK. “On 30 September 2026, law enforcement took control of KillSec’s leak site, securing at least 110 terabytes of data against further unauthorised access.” reads the press release published by EUROPOL. “KillSec stole sensitive data by exploiting vulnerabilities and poorly secured access points to organisations’ systems. Around 500 of the suspected attacks have so far been identified as successful. This figure may change as investigators examine the evidence seized during the operation.” – https://securityaffairs.com/200200/cyber-crime/operation-killswitch-police-dismantle-killsec-ransomware-group.html
Mississippi mayor says ransomware incident led city to shut down systems – A ransomware attack has shut down the computer systems of Vicksburg, Mississippi, the city’s mayor told residents on Thursday evening. Mayor Willis Thompson published a statement in the local newspaper saying the city is investigating a ransomware attack that has not impacted emergency services but has affected payments for utilities. He said the system shutdown was “temporary”. Thompson said no one’s services will be shut off while the investigation is ongoing and no penalties will be issued for late payments. He later told the Vicksburg Post that the city has been working on the recovery effort with the FBI, Department of Homeland Security and other state officials alongside private cybersecurity experts. “One of the top priorities of our investigation is determining whether there was any compromise of personal or confidential information relating to current and/or former customers, contractors, vendors, employees, or affiliated business partners of the City,” Thompson said in a statement. – https://therecord.media/vicksburg-mississippi-government-ransomware-attack
Delhi approves 2,500 AI-enabled cameras for traffic monitoring and enforcement – The Cabinet Committee on Economic Affairs (CCEA) in India approved a ₹1,789.52 crore Intelligent Traffic Management System (ITMS) for Delhi on 30 September, expanding the use of AI-assisted monitoring and automated enforcement across the capital’s road network. The project, which will be operated by Delhi Police, will cover 42 traffic corridors and be introduced in three phases over 24 months, followed by five years of operation and maintenance. According to the government, it will combine real time traffic data, adaptive traffic signals, automated enforcement and integrated command and control systems. The official Cabinet announcement says the system will cover 1,092 Adaptive Traffic Control System locations and 1,029 enforcement locations. Automated systems will detect offences including speeding, red light and stop line violations, wrong way driving, illegal parking, failure to wear helmets or seat belts and mobile phone use while driving. – https://dig.watch/updates/delhi-ai-cameras-traffic-monitoring-enforcement
Intelligence, Defence, Military, Warfare
In the Online Age, PRC Foreign Intelligence Evolves – Chinese intelligence efforts are increasingly moving away from human intelligence (HUMINT) collection to digital operations launched from within the People’s Republic of China (PRC). The Ministry of State Security (MSS), military intelligence, and others have started to employ online recruiting to replace face-to-face contact for foreign assets. Beijing likely adopted the new model in response to the 2018 arrest of Xu Yanjun—the first MSS officer to be tried and convicted in the United States—which highlighted counterintelligence risks facing officers operating outside PRC borders. The MSS is using LinkedIn en masse to recruit and develop assets overseas. These operations frequently involve headhunter-style approaches and commercial pretexts to collect resumes and gather intelligence. Traditional HUMINT remains an important component of MSS tradecraft. MSS agents have started to integrate face-to-face contact in the PRC for managing their overseas contacts while their assets abroad deploy digital tools to facilitate operations. – https://jamestown.org/in-the-online-age-prc-foreign-intelligence-evolves/
In Ukraine, data centers ‘facing increased attacks’ in recent weeks: Ex-official – Faced with a recent surge in Russian attacks on its data centers, Ukraine must urgently build more underground networks and boost its anti-ballistic and counter-drone capabilities to avoid the loss of critical infrastructure, the country’s former deputy minister of defense has warned. In the last month, Moscow has continued to strike Ukrainian data centers and communications infrastructure, which were not prime targets earlier in the war. On Sept. 27, the headquarters of the war-torn country’s central mobile service provider, Kyivstar, was hit by a Russian drone strike. The same day, the Russian Defense Ministry also claimed responsibility for targeting the data center of Vodafone Ukraine, the country’s second-biggest cell provider network, according to Reuters. “Data centers have been facing increased attacks these last few weeks, including strikes by Shahed drones. Russia also increased its cyberattacks even before the war started and has been adapting them to be even more sophisticated since,” Kateryna Chernohorenko, who was deputy minister of defense for digital transformation of Ukraine from 2023 to 2025, told Breaking Defense in an interview Sept. 26 at the IT Arena conference in Lviv. – https://breakingdefense.com/2026/10/in-ukraine-data-centers-facing-increased-attacks-in-recent-weeks-ex-official/
Ex-defense minister’s futuristic ‘Army of Robots’ raises eyebrows in Ukraine – Ukraine’s former Minister of Defense Mykhailo Fedorov has launched a new project to build a robotic army, including humanoid soldiers, intended to largely replace soldiers on the battlefield. He contends the initiative will “save lives,” but it has also sparked mixed reactions at the IT Arena conference here and beyond. During the conference, which took place here from Sept. 25-27, Fedorov, who was controversially dismissed in July by Ukraine’s President Volodymyr Zelenskyy, unveiled an envisioned “Army of Robots” with a sleek, CGI video showing humanoid robots aiming rifles, fighting in trenches, and marching in a massive formation. “In 2022, we launched the Army of Drones. […] Today, drones account for more than 95% of strikes,” Fedorov said in a social media post on Sept. 26. “But technological warfare does not stand still. We need the next breakthrough. Our bet is on the robotization of warfare. The mission of the Army of Robots is to save lives”. During his conference speech, he explained that the future battlefield he envisions would be one where sensors detect threats first, followed by the deployment of robots, based on the belief that robotic platforms “should fight instead of people”. Fedorov announced in August that he was launching his own defense company, with few details about what it would entail, except that its first major investor would be Palantir CEO Alex Karp. In his speech here, Fedorov did not say what role his company would play, but emphasized that his vision would not be supported by one single company anyway. Rather, he saw an “ecosystem” that will involve investments in defense companies, R&D capabilities, testing alongside the military and scaling of effective systems. – https://breakingdefense.com/2026/10/ex-defense-ministers-futuristic-army-of-robots-raises-eyebrows-in-ukraine/
French Air Force plans loyal wingman flight in 2028 in sovereign AI push– France’s Air Force, together with the country’s Directorate General for Armament and the agency for defense artificial intelligence, plans to fly crewed fighter jets in collaboration with combat drones in 2028, part of a project to experiment with AI in combat aviation dubbed Hypairion. The project seeks to develop open, modular architecture for combat aircraft and sovereign AI technology that gives the government greater control over the tech, rather than having to rely on companies for critical systems, said Col. Cédric, head of the Air Force capability-development office, in a press briefing here on Thursday. As AI becomes more powerful, the likes of Airbus and BAE Systems in Europe and General Atomics and Anduril in the United States are leveraging the technology to develop collaborative combat aircraft, also referred to as “loyal wingmen.” While many of those projects are company-driven, the French project would keep some of the critical technology under state control. “What Hypairion is really about is equipping the Air Force and our platforms with architectures that are open, modular and owned by the government,” Col. Cédric said. He was identified only by his first name, in line with the Armed Forces Ministry’s policy for military staff. “Our goal is to move away from proprietary systems that lock down sensors and weaponry”. The Ministerial Agency for Defence AI will be in charge of providing sovereign AI technology for the project in several areas, particularly command and control as well as mission autonomy, said Col. Thibault, deputy director for AI applications at the agency, known as Amiad. He described mission autonomy as AI that assists pilots with mission planning, in-flight decisions and post-mission analysis. – https://www.defensenews.com/global/europe/2026/10/02/french-air-force-plans-loyal-wingman-flight-in-2028-in-sovereign-ai-push/