AI Law Daily Briefing — 11 July 2026

Digital Omnibus on AI enters into force — Annex III delay to Dec 2027 now binding law

The Digital Omnibus on AI entered into force this week following Official Journal publication, capping the European Parliament's 16 June vote and the Council's 29 June final green light. The Annex III high-risk deadline (2 Dec 2027) and Annex I deadline (2 Aug 2028) are no longer provisional — they are now legally binding, so the 'not yet law' caveat on those dates can come off the tracker. Separately, GPAI penalty enforcement activates 2 August 2026 and applies retroactively to violations dating back to August 2025 — any foundation-model provider treating the first year as a free pass should assume past conduct is now exposed. Article 50 transparency obligations are unaffected by the high-risk delay and remain enforceable 2 August 2026 as originally scheduled.

Source: Tech Times

Correction, added 25 July 2026: this briefing stated the Digital Omnibus had entered into force as of 11 July 2026. That was premature — the source article was itself wrong on timing. Regulation (EU) 2026/1744 was not published in the Official Journal until 24 July 2026 and did not enter into force until 27 July 2026. The Annex III (2 Dec 2027) and Annex I (2 Aug 2028) deadlines described above are correct; the "entered into force this week" claim in the paragraph above was not accurate at the time of writing.

Tinubu directs FCCPC to investigate Big Tech and generative AI platforms over news scraping

President Tinubu has directed the FCCPC to investigate Meta, Alphabet, X and unnamed generative AI platforms over alleged unauthorised scraping of Nigerian news content for model training, following a joint petition from the Nigerian Press Organisation. This runs alongside — not instead of — the broader NITDA/NCC/NDPC digital-platform rules postponement reported 3 July: the FCCPC action targets AI training-data practices specifically under the FCCPA 2018, a competition/consumer-protection statute, not the stalled AI-specific governance framework. For any client operating AI products with Nigerian user bases or training data sourced from Nigerian media, this is a live enforcement track independent of the paused AI regulations — FCCPC has form here, having fined Meta $220M in 2025 (under appeal).

Source: Federal Competition & Consumer Protection Commission (Nigeria)

EU Parliament passes Chat Control 1.0 by default — AI-based message scanning legal to 2028

EU Parliament passed Chat Control 1.0 into law on 9 July without a majority of MEPs ever voting for it — a second-reading procedural quirk meant the 314 votes against fell 47 short of the 361 needed to block the Council's text. The law authorises (does not mandate) voluntary scanning of unencrypted messaging and email services using AI image classifiers and grooming-detection text analysis, running until April 2028; end-to-end encrypted services such as WhatsApp and Signal stay outside scope. Relevant to any client running AI-based content moderation or trust-and-safety tooling on EU users' communications — coverage citing the Commission's own implementation data puts the AI classifiers' false-positive rate as high as 20%, a proportionality question likely headed to the CJEU.

Source: Tech Times