AI News from September-October 2026

Published
16 Oct 2026
Read time
9 min read
Category

Recent AI Laws and Regulations

Australia to introduce safety requirements for AI companies

Australia’s Science, Technology and the Digital Economy Assistant Minister Andrew Charlton said the government will introduce obligations for AI companies that would require organizations to follow similar testing protocols to banking and aviation laws.  Under the proposed law, AI companies would have to prove they are properly testing models and preventing harm. More here.

The EU’s First AI Standard is here

The new standard on AI quality management systems has been published in the Official Journal of the European Union. The standard EN 18286:2026, “Artificial intelligence – Quality management system for EU AI Act regulatory purposes,” specifies requirements and provides guidance for defining, implementing, and maintaining a quality management system for organizations that provide high-risk AI systems.  More here.

US government task force to develop civil, criminal standards for AI developers

The Director of the US Office of Personnel Management Scott Kupor said the government’s “super intelligence” task force will work alongside the Department of Justice and other law enforcement agencies to hold AI developers to their safety commitments. The task force will focus primarily on establishing criminal and civil standards for developers based on existing law. More here.

US Senator Cantwell unveils 6-point AI governance policy proposal

US Senate Committee on Commerce, Science and Transportation Ranking Member Maria Cantwell, D-Wash., released an AI governance policy framework that addresses needs for “safety standards, continuous testing, and reporting of serious failures.” The framework is built around six principles, including “enforced standards,” independent testing, transparency and accountability, and safeguards against social harms.  More here.

New York City Council introduces bills to establish liability for AI-related risks

The New York City Council proposed a number of bills to establish AI standards, aiming to advance consumer protections and increase liability for companies whose tools could have critical safety risks. City Council Speaker Julie Menin said while the U.S. has not enacted federal AI safeguards, “New York City will explore nation-leading measures that protect the public while allowing innovation to thrive. We can and must be both pro-innovation and pro-safety. More here.

AI Legal Cases This Month

Getty Images loses copyright case

Getty Images has lost its UK copyright case against Stability AI over the training and operation of Stable Diffusion, the company’s text-to-image generator. Getty alleged that Stability AI used its stock images without permission to train the model, but dropped its primary infringement claim because it could not prove the relevant training occurred in the UK. Its remaining secondary infringement claim also failed. Although the High Court accepted that an AI model stored in the cloud could be an “article” under the Copyright, Designs and Patents Act, it held that an “infringing copy” must itself reproduce the protected work. Because Stable Diffusion did not retain reproductions of Getty’s images, the model was not an infringing copy—a significant distinction between a system developed using copyrighted material and one that actually stores or reproduces it. More here.

European Commission launches AI Act probe into Anthropic, OpenAI

The European Commission has confirmed the AI Office will open investigations into recent cybersecurity incidents tied to autonomous AI agents belonging to OpenAI and Anthropic under the EU AI Act, Euronews reports. The investigations could become precedent-setting under the AI Act; however, some legal experts dispute the Commission has the power to investigate because the incidents happened during the respective companies’ testing phases before the models were available in the EU. More here.

Class-action lawsuit claims OpenAI evades parental safety measures, addicts children

An unnamed mother of two adolescents sued OpenAI in a US class-action lawsuit alleging ChatGPT evades parental consent tools and addicts children with its features to steal their personal information for model training, Courthouse News Service reports. “(ChatGPT) wrongfully has extracted vast amounts of highly sensitive and valuable data from millions of children,” the complaint stated. “(OpenAI) then used that priceless trove of children’s information to train large language models, build detailed profiles of children and enrich itself.” The filing alleges children only need to self-identify they are over age 13 to use ChatGPT. More here.

FTC reportedly has open probe into AI developers

The New York Post reports the US Federal Trade Commission is preparing to issue civil investigative demands and compel testimony from the leading AI firms about the risks and potential rogue capabilities of their frontier models. An FTC official said the probe began “a few weeks ago” before OpenAI’s Hugging Face hack announcement. They added the agency’s objective is to learn “about the dangers they allege their products may have to consumers, to Americans.” The FTC has not publicly announced the probe. More here.

Character.ai files for bankruptcy after claims its chatbot led to mental health crises in teens

AI startup Character.ai filed for Chapter 11 bankruptcy after facing legal claims that its app contributed to mental health crises and death by suicide among adolescents, CNN reports. The company plans to request a court-supervised sale by the end of the year. Disney has agreed to purchase some of the company’s intellectual properties for USD20 million. More here.

OpenAI engineer bugged exec meeting room

US federal prosecutors have charged OpenAI engineer Abdulrahman Alfozan, a Saudi national, with trying to eavesdrop on company leaders via a recording device hidden in its SF office, and OpenAI says its own AI helped trace the device to him.  More here.

Anthropic seeks Full Appeals Court Review of Pentagon Ban

Anthropic filed an anticipated appeal in its ongoing fight against a blacklisting by the U.S. Department of Defense, arguing that the case’s far-reaching implications demand a larger panel of judges to decide. A request for an en banc hearing is an expected stage in the appeals process. More here.

Security News This Month

OpenAI breaches Medicare and other sites

OpenAI agents hacked into a ‘Medicare portal’ while supposedly looking for health data for a sandboxed training task. OpenAI then waited 3 months to send an email to the general Services Australia inbox. More here. The Albanese government is said to be “urgently investigating” the hack as well as looking into the country’s cyber defences and deciding whether any laws need to be strengthened.  A former Data61 senior programmer warned ministers in April that sloppy record-keeping would lead to a breach – here. However, some cybersecurity experts and tech investors think the Albanese government has blown the Medicare breach out of proportion, arguing that “the website’s code was set up to send data to anyone with the skills to request it”.  More here.

Australian breach tip of the iceberg

Adding to the concerns over AI, Axios reported over the weekend that “OpenAI, Anthropic and security researchers are investigating tens of thousands of incidents in which their frontier models took steps that outside evaluators would consider problematic.”

OpenAI admits to another government breach after PM lashing

OpenAI has admitted that one of its AI models accessed additional non-public Australian government data, just days after Prime Minister Anthony Albanese lashed the company for taking weeks to disclose that its agents hacked Medicare. More here.

Canada and UK announce renewal of AI security partnership

The UK reaffirmed their partnership on AI security during a visit to London by Canada’s Minister of Artificial Intelligence and Digital Innovation Evan Solomon. The pact focuses on strengthening information sharing on frontier AI risks and capabilities, best practices for evaluation and developing joint technical standards. The partnership will also facilitate increased cooperation between the Canadian AI Safety Institute and the UK AI Security Institute.  More here.

Reports find cybercriminal use AI in 84% of ransomware attacks

According to the 2026 Ransomware Risk Report compiled by the cybersecurity firm Semperis, cybercriminals used AI in 84% of attacks, CBS News reports. The report also found that hackers exploited victims’ own AI systems in 44% of ransomware attacks. The report polled 1,200 IT and security professionals across nine countries and eight industrial sectors. In total, 87% of victims experienced attacks that targeted their identity systems or the IT infrastructure controlling their network. More here.

Australia’s AI Parliamentary Investigations

2026 has seen a raft of parliamentary AI inquiries across federal and state jurisdictions. Inquiry into Security Breaches and AI Agent Conduct (Ad-Hoc Executive/Parliamentary Review) is the urgent inquiry into the immediate aftermath of the Medicare hack becoming public knowledge and will seek to understand if OpenAI can be charged with any breaches of the law. Other inquiries include a Joint select committee on artificial intelligence, set to report on November 30, its scope includes opportunities arising from AI to lift productivity, growth, economic competitiveness, and living standards. More here. NSW Legislative Council public accountability and works committee: Inquiry into the Development and Regulation of Data Centres was established at the end of January and is looking into the development and sustainability of data centre development in New South Wales.  More here.

Researchers make ‘first contact’ with a rogue AI fleet

A research group called the “swarmchasers” just revealed it traded messages with a fleet of rogue Tencent AI agents, in what it believes is a “first contact” moment, after catching them trying to scrape Amap, Alibaba’s version of Google Maps. More here.

Indeed says AI is boosting jobs

An Indeed Hiring Lab analysis challenges fears that generative AI will erode knowledge work and wages, finding that advertised pay is rising fastest in highly exposed US occupations such as software development and analytics. About a quarter of listed jobs are highly exposed to AI, 54% are moderately exposed, while roles including nursing and food preparation have minimal exposure.  More here.

Heidi has big capital raising

One of the biggest AI Scribe companies in the world, Heidi, announced it has raised an eye-watering $US340 million to add AI Agents to its medical software. Digital Right Watch’s  report on AI Scribes, found these tools exist in a regulatory grey-zone and cause real harm.

Global researchers call for international body to oversee advanced AI

A coalition of 400 researchers wrote an open letter urging governments to work together to establish a global AI safety agency, Reuters reports. The researchers, including Geoffrey Hinton and Yoshua Bengio, noted with technology advancements raising concerns about consumer safety, an international regulatory body could work to ensure users are protected by implementing global safety, transparency and testing standards. More here.

Ready to turn insight into action?
Connect with Privacy 108.

"*" indicates required fields

This field is for validation purposes and should be left unchanged.
Privacy 108 collects your name and contact details to respond to your enquiry and communicate with you about it. If you do not provide this information, we may be unable to respond. We do not disclose this information to third parties. For more information about how we handle your personal information, including how to access or correct it or make a complaint, please see our Privacy Policy or contact us at hello@privacy108.com.au.
Related articles
Subscribe to our Newsletter

"*" indicates required fields

This field is for validation purposes and should be left unchanged.

Privacy 108 collects your name and email to send you our newsletter. If you do not provide this information, we will be unable to send it to you. We may use third-party service providers (such as email marketing platforms) to distribute our communications. Some providers may store information overseas, including in the United States. For more information about how we handle your personal information, including how to access or correct it or make a complaint, please see our Privacy Policy or contact us at hello@privacy108.com.au. You can unsubscribe at any time using the link in our emails or by contacting hello@privacy108.com.au.