Week of 18 May 2026
Trump and Xi placed AI safety on the Beijing summit agenda, with Treasury-led bilateral dialogue being discussed.
Key points
- Talks focus on access controls, best practices for advanced models, and limiting non-state actor access - not a binding treaty.
- Outcome mechanisms, if formalised, could reshape export controls, chip access, and frontier-model procurement conditions globally.
Trump is expected to sign an executive order creating a voluntary pre-release AI disclosure framework for US government and critical infrastructure providers.
Key points
- The 90-day pre-public model access window sets a US precedent that could influence Australian pre-deployment safety assessment expectations.
- The framework is voluntary, limiting its direct regulatory force - Australian agencies should note this distinction when tracking US AI governance signals.
The European Commission has released draft guidelines clarifying which AI systems qualify as high-risk under the EU AI Act.
Key points
- Stakeholder feedback is open until 23 June 2026 - Australian AI providers operating in EU markets may be directly affected.
- Guidelines include practical examples to help providers and deployers self-assess high-risk classification obligations.
EU AI Office held third-round stakeholder meetings to finalise the Code of Practice on AI-Generated Content transparency.
Key points
- The final draft covering marking, watermarking, deepfake disclosure, and labelling obligations is expected in early June 2026.
- Debates centre on mandatory versus voluntary measures and compliance burden - tensions likely to recur in any Australian equivalent framework.
GoTo-commissioned survey of 2,500 global workers finds 39% report AI use has weakened their skill sets.
Key points
- Nearly one in four IT leaders report AI-related mistakes have already affected customers or the bottom line.
- Survey is vendor-commissioned and measures self-reported perceptions, not objective skill decline - treat with appropriate caution.
OECD AI blog addresses shared foundations for collective AI security across member nations.
Key points
- Covers prompt injection, AI agents, and model poisoning - security risks relevant to Australian government AI deployments.
- Extracted text is minimal; full substance of the piece is not available for detailed assessment.
AI Now Institute is launching a dedicated research portfolio examining AI deployment risks across US healthcare systems.
Key points
- Key concerns include patient safety failures, workforce displacement, regulatory gaps, and corporate consolidation - relevant to Australian health AI governance debates.
- Focus is US-specific; Australian health AI governance context differs, limiting direct applicability for APS readers.
California Governor Newsom signed an executive order directing state agencies to prepare for AI-driven workforce disruption.
Key points
- The order tasks procurement agencies to develop AI vendor certification rules within 120 days, including watermarking and bias safeguards.
- This is a US state-level development with no direct Australian regulatory parallel, though procurement parallels are worth noting.
HCLTech survey of 467 G2K executives finds 24-43% of major AI initiatives expected to fail (figures conflict across sources).
Key points
- 76% of surveyed executives say Responsible AI concerns have delayed deployments - a tension familiar to APS agencies.
- Private-sector vendor survey with methodological inconsistencies; limited direct applicability to Australian government context.
Google and OpenAI are shifting investment toward general agentic AI scientists rather than specialised scientific tools.
Key points
- OpenAI's general-purpose reasoning model independently disproved a mathematics conjecture, signalling genuine research capability.
- Limited direct relevance to APS AI governance or procurement - included as a horizon-scanning signal.
Anduril and Meta are jointly developing AI-powered smart glasses for US Army combat use, including threat identification and strike recommendation.
Key points
- AI-enabled military wearables raise governance questions about autonomous decision-support and human oversight in lethal contexts.
- Limited direct relevance to Australian federal agencies; useful context for defence-AI and autonomous systems policy discussions.
NIST NCCoE is hosting a June 9 webinar on Privacy-Enhancing Technologies testbed and Dioptra AI security platform.
Key points
- Work focuses on securing AI model training on sensitive genomic data using differential privacy and federated learning.
- Niche technical event with limited direct APS applicability; useful context for AI privacy and security practitioners.
HPE Threat Labs found governments were the most frequently targeted sector globally in 2025.
Key points
- AI-augmented cyber threats are a real and growing concern, but this article is primarily vendor-positioned content.
- Limited direct APS governance or policy signal - included for contextual awareness only.
The European Commission is reviewing its 2019 Copyright Directive and seeking stakeholder views on generative AI licensing challenges.
Key points
- AI's intersection with copyright is one of several review threads - not the sole or primary focus of this consultation.
- Limited direct relevance to APS practitioners; EU copyright law does not bind Australian agencies.
The EU Commission held a roundtable to launch the first vetted researcher data access requests under the DSA.
Key points
- AI features on platforms are among the research focus areas flagged in 49 applications received so far.
- Limited direct relevance to APS; included as context for EU platform accountability developments.
MIT Technology Review video roundtable covers the Musk v. Altman trial and its implications for the AI industry.
Key points
- The trial concerns OpenAI's nonprofit-to-for-profit conversion - a governance dispute with broader AI sector implications.
- Limited direct relevance to Australian federal agencies; useful background on US AI sector governance disputes.
A US jury ruled against Elon Musk in his lawsuit over OpenAI's conversion from nonprofit to for-profit.
Key points
- The case turned on statute-of-limitations technicalities rather than substantive AI governance principles.
- Limited direct relevance for Australian federal agencies - included for broader AI-sector context only.
A federal jury dismissed Musk's $150 billion lawsuit against OpenAI on statute of limitations grounds.
Key points
- The verdict clears a legal obstacle to OpenAI's potential IPO, with commentators citing a near-$1 trillion valuation.
- Limited direct relevance to APS work; context only for those tracking AI sector governance and commercialisation trends.
An unnamed leadership summit called for sustained dialogue on AI's role in education and healthcare.
Key points
- No specific country, agency, policy commitment, or actionable outcome is identified in the item.
- Extremely low signal for APS readers - vague summit coverage with no substantive detail or Australian angle.
Week of 11 May 2026
Good Ancestors' May 2026 newsletter covers biosecurity-AI risk, Australia's AI strategy, Mythos cyberattack capability, CAISI testing agreements, and DTA policy.
Key points
- DTA's Policy for the Responsible Use of AI in Government v2.0 is now mandatory; AI use-case registers due across non-corporate Commonwealth entities by mid-2026.
- Australia is excluded from Anthropic's Project Glasswing defensive coalition; frontier AI cyber risk to critical infrastructure has no current Australian mitigation mechanism.
AI chatbots including Gemini and ChatGPT are exposing real personal phone numbers drawn from training data.
Key points
- DeleteMe reports a 400% rise in customer queries specifically referencing generative AI tools exposing personal data.
- PII leakage from LLMs is directly relevant to APS obligations under the Privacy Act and responsible AI policy.
Amazon employees gamed internal AI usage metrics by automating token consumption via an agent platform called MeshClaw.
Key points
- Illustrates a governance failure: raw consumption metrics as AI adoption KPIs create perverse incentives over genuine productivity gains.
- Security concerns arose from agents running with broad permissions on employee hardware - a least-privilege governance gap.
Alan Turing Institute research links sustainability measures in Defence AI procurement to increased force resilience.
Key points
- Findings are UK-focused but offer transferable framing for Australian Defence AI governance and procurement policy.
- Extracted text is truncated - full substance of the research recommendations is not available from this item.
Australia's Social Media Minimum Age Act commenced December 2025; early implementation reveals inconsistent platform compliance.
Key points
- AI-driven age assurance technology is proven viable; the gap is governance, enforcement, and operational consistency.
- APS-relevant assurance lessons - independent testing, lifecycle controls, conformity assessment - apply broadly to outcome-based regulation.
A Chinese think tank representative privately requested access to Anthropic's Mythos model at a Singapore meeting; Anthropic refused.
Key points
- US National Security Council officials were alerted and reacted with concern, signalling frontier AI access controls as a live geopolitical issue.
- No technical details about Mythos have been disclosed; the governance significance outweighs the technical content of this report.