17 articles
Index of editorial tags and filtered coverage.
OpenAI will begin adding an invisible watermark, called textGrain, to text generated by ChatGPT and Codex for users in the European Union to comply with the AI Act's transparency rules; on the API the feature is opt-in and off by default worldwide.
From 2 August 2026, the European Commission, via the AI Office, is the exclusive supervisor of general-purpose AI (GPAI) model obligations, with fines up to €15 million or 3% of worldwide turnover. 2 December expands Article 50 transparency and intimate-image prohibitions.
On 1 October 2026, EU Perspectives analysed how ChatGPT—designated a Very Large Online Search Engine (VLOSE) under the DSA after reporting about 159.1 million average monthly EU users—now stacks DSA and AI Act duties, with the proposed KIDS Act as a potential third layer on chatbots and children.
On 30 September 2026 Euronews contrasted the White House Joint Commitment—“morally binding,” with no legal force—with Europe’s AI Act: since 2 August the AI Office can demand information, evaluate models, restrict market access and fine up to €15 million or 3% of global turnover. Same signatories, two compliance philosophies.
The Digital Omnibus opened a four-month window for generative systems already on the market before 2 August 2026; other Article 50 duties have been enforceable since that date. The European AI Board’s 9th plenary (17 Sep) offered no further relief: fines up to €15M or 3% of turnover, with detector interoperability aimed at 2 February 2027.
Since 2 August 2026 Anthropic embeds SynthID-Text-style invisible text watermarks in new Claude models, with detector access on request, C2PA on generated images/files, and coverage across API, Claude Code, AWS, and GCP—aligned with AI Act Article 50 transparency duties.
Confirmed 1 September 2026: the European Commission opened information requests to 30+ AI providers under GPAI powers live since 2 August; no AI Act fine has issued yet.
As of 23 September 2026, Article 50 transparency has applied since 2 August; machine-readable marking grace for pre-August generative systems ends 2 December. The Digital Omnibus did not freeze Art. 50; it deferred Annex III high-risk items.
From 1–2 September 2026 the AI Office is sending formal information requests to more than 30 providers on advanced-model safety, copyright and transparency; incomplete answers can trigger fines up to €15m or 3% of global turnover.
On 22 September 2026 at Atlanta’s Transformation Excellence Summit, SAP integrated Signavio, LeanIX, WalkMe and Cloud ALM with an AI Agent Hub, an AI Act/NIST Governance Assistant and production agent monitoring.
On 21 September, the EU debate links Claude Opus containment failures, OpenAI agents and the fine regime (up to 3% or €15M since 2 Aug 2026); Project Glasswing and ENISA Mythos access backdrop the fight over reporting thresholds.
A letter from Brando Benifei and Axel Voss, reported by MLex around September 17, 2026, urges the Commission to open Digital Europe funding so the European Data Protection Supervisor can meet AI Act tasks.
September 2026 analysis of AI Act Article 50 transparency, the June/July Code of Practice, and how Anthropic marks Claude text and images (SynthID-Text/C2PA) globally since about August 2.
Around September 16, 2026, InfoHandle reported the AI Act’s first coordinated high-risk list: border biometrics, hiring video scoring, non-bank credit, and rail signaling, with FRIAs due 15 Jan 2027, letters to OpenAI/Mistral/Google, and Q1 2027 audits.
In late August and early September 2026 the European Commission, with Henna Virkkunen leading the digital dossier, launched first information requests to 30+ general-purpose model companies under the AI Act.
The European regulation began compliance timelines for bans, transparency, and general-purpose models.
In December 2023, European negotiators reached political agreement on the AI Act, with risk-tiered obligations.