EU Council Approves AI Act Omnibus Reforms, Prompting Compliance Shifts Across Europe

The Council of the European Union has given its final approval to the AI Omnibus package, a set of amendments to the EU AI Act. The reforms postpone the start of the high‑risk regime to 2 December 2027 for stand‑alone systems and to 2 August 2028 for high‑risk AI embedded in products. They also introduce a new prohibited practice banning AI‑generated non‑consensual sexual imagery and child sexual abuse material from December 2026, shorten the transparency‑solution grace period to three months, and push back the deadline for national AI sandboxes to August 2027. The package clarifies how the AI Act interacts with sector‑specific legislation, allowing the Commission to limit overlapping obligations.

European ICT service providers are being urged to verify that any AI functionality they develop or integrate complies with the Act. Responsibility for compliance rests primarily with the client, who must ensure that high‑risk applications—such as those used in recruitment, medical diagnostics, or critical infrastructure—meet documentation, risk‑assessment and audit requirements. Providers may refuse work that cannot be brought into conformity.

Critics argue the EU framework does not fit African realities. In Kenya, regulators lack the technical capacity to audit complex models, and data sources are fragmented and informal. Observers call for proportionate, locally‑designed rules—including basic audit logs and human‑appeal mechanisms—rather than direct import of EU standards.