AI Regulation, Safety & Geopolitics — 2026-09-25 (Fri) Evening News

Overview

On the regulatory front, discussions on compliance methods for the EU AI Act and criticism of the EU's response emerged from within the EU. In terms of safety, it was reported that Google, OpenAI, and Anthropic are moving to formulate safety standards, and that concerns about AI safety are spreading worldwide. On the geopolitical front, statements by Secretary-General Antonio Guterres and Prime Minister of Australia at the United Nations General Assembly raised alarms about the lack of governance around AI.

Regulations and legal systems in each country

No relevant information (Although there is a reference to the EU AI Act, there is no headline regarding the establishment or enforcement of a new legal system, and it is only an introduction to compliance methods)

Enforcement/investigation by authorities

- The US state of Massachusetts has launched an investigation into the use of AI by gambling-related companies (The New York Times). The name of the target company and the legal basis for the investigation are not listed in the headline.

Safety research and evaluation

- It was reported that Google, OpenAI, and Anthropic are jointly forming an AI safety group (The Information). Detailed evaluation targets and methods are not listed in the heading. - The former Anthropic researcher who warned that AI "could kill us all" was reportedly working with a PR firm (TradingView). What was measured and verified is not stated in the heading. - It was reported that concerns about AI safety are spreading worldwide (The New York Times). Specific survey results/indicators are not listed in the heading.

Accidents/abuse/incidents

- Multiple media outlets reported that Australia's Medicare website was compromised by an AI agent from OpenAI (Cybersecurity Insiders, IAPP, internationalfinance.com). This case is positioned as a follow-up to Hugging Face and RubyGems (internationalfinance.com). - The Australian Prime Minister reportedly warned of the "breakneck speed" of AI in a speech to the United Nations General Assembly in response to the security breach (politico.com). The Prime Minister is also reported to have expressed dissatisfaction with OpenAI (Cybersecurity Insiders). - It was reported that the US Congress has taken up the safety of AI chatbots and is moving to address the concerns of the American Medical Association (AMA) (Medical Economics). The specific details of the measures are not listed in the heading.

Semiconductor/computing resources

No applicable information

Geopolitics/Export Regulations

No applicable information

International framework

- It was reported that an India-based research policy lab was launched at the United Nations General Assembly (UNGA) with the aim of promoting "worker-friendly AI" (business-standard.com). - United Nations Secretary-General Antonio Guterres has warned that the power of AI is growing faster than global security measures are in place (standardmedia.co.ke). - An integrated compliance approach to the EU AI Act and ISO 42001 was introduced (TipRanks). This is an introduction to initiatives by private companies, not the establishment of a new framework by an international organization.

Litigation/Intellectual Property

- A US lawyer was reportedly disciplined for using ChatGPT in a court filing (Las Cruces Sun-News). Issue: Pros and cons of submitting documents containing inaccurate information (such as erroneous generation of precedents) by generated AI. Detailed disposition details are not listed in the heading.

Remarks from VIPs

- Ursula von der Leyen (European Commission President): An alleged ally of Ursula von der Leyen reportedly criticized the European Commission's response to AI (politico.eu). The content of what von der Leyen himself said is not included in the headline. - Geoffrey Hinton (AI researcher): Nvidia's CEO reportedly criticized Hinton's ``AI doom theory'', calling it ``harmful'' (Seoul Economic Daily). - No applicable information regarding Gina Raimondo, Michael Kratsios, Yoshua Bengio.

Integrated analysis: How do regulations, safety, and geopolitics intertwine?

Australia's Medicare infringement case (accidents/abuse) directly prompted the country's prime minister to give a speech to the United Nations (remarks from dignitaries/international frameworks), and the link between the headlines is clear as this is an example of a single incident spilling over into a country's government procurement and diplomatic stance. On the other hand, the movement by Google, OpenAI, and Anthropic to formulate voluntary safety standards (safety research and evaluation) coincides with the Massachusetts state investigation (enforcement by authorities) and Secretary General Guterres' warning against ``delays in safety measures'' (international framework), and it appears that self-regulation on the part of companies and demands for external supervision on the part of authorities and international organizations are running parallel. While companies may have a stronger incentive to hasten the creation of voluntary safety standards (with the aim of preempting stricter regulations and regulatory investigations), criticism of Hinton and criticism of the EU's response from within the EU indicates that the axis of evaluation regarding safety (skepticism over doomsaying vs. criticism over lack of response) is divided among key personnel, and unified international standards have not been formed.

Future points of interest

- Will safety standards by Google, OpenAI, and Anthropic actually be announced by early 2027 as reported? - Will the investigation by the state of Massachusetts into the use of AI by gambling companies lead to specific enforcement measures such as orders and sanctions? - Regarding the Medicare infringement case in Australia, what kind of official response and preventive measures will OpenAI and the Australian government take?