AI Updates
The AI developments that matter, straight from the source
No rushed headlines, no hype — just the developments actually worth your attention, summarized in short paragraphs with a direct link to the official source for every item.
We rely only on official sources (the EU, NIST, OECD, or direct company announcements) — never on unverified news-aggregator sites. This page is updated periodically.
August 2025 — in force now
Governance · EU
General-purpose AI (GPAI) obligations under the EU AI Act took effect
Since 2 August 2025, providers of general-purpose AI models (such as large language models) must keep detailed technical documentation, adopt a copyright policy, and publish a summary of the content used to train the model. The most impactful models ("systemic risk") face additional obligations: notifying the European Commission, assessing and mitigating risks, and reporting serious incidents.
Source: European Commission ←
April 2026
Risk Management · NIST
NIST extends its AI Risk Management Framework to critical infrastructure
The US National Institute of Standards and Technology (NIST) released a concept note for a new profile guiding critical-infrastructure operators (energy, water, transport) on managing AI-system risks. This comes as part of a broader revision of the core AI RMF under the White House AI Action Plan.
Source: NIST ←
2026
Labour Market · OECD
OECD Employment Outlook 2026: AI is reshaping, not eliminating, the job market — unevenly by region
The report finds that AI's and other technology's effects on jobs vary sharply by region: some regions lose manufacturing jobs while others gain service and non-routine work. It stresses that relocation alone doesn't close labour-market gaps, and calls for integrated, place-based policies combining regional and employment support.
Source: OECD ←
Get a free weekly summary
One best analysis or update, straight to your inbox every week. No spam, unsubscribe anytime.