RESOURCE · EU AI ACT MAPPING
This mapping shows where Mergetic capabilities can support operationalisation of Articles 9, 12, 14, 50 and 72. It does not certify compliance, replace legal analysis or decide whether the Act applies to a deployment.
THE CURRENT TIMELINE
Article 50 transparency duties generally began applying on 2 August 2026. A transition to 2 December 2026 applies to specific pre-existing systems covered by Article 50(2). Regulation (EU) 2026/1744 deferred the Chapter III high-risk requirements for Annex III systems to 2 December 2027 and Annex I systems to 2 August 2028. Article 72 was not included in those Chapter III deferrals. Applicability still depends on the system, role and facts.
Article 50
Capability: For text your customers will see, Mergetic checks that the required "this was produced by AI" notice is present before release. If it's missing, the answer goes back to be fixed or is held for a person, and the result is recorded.
Boundary: Text-output capability support only. Mergetic is not a synthetic-media marking or watermarking tool, and this is not a compliance determination.
Article 9
Capability: Your risk rules are applied to each decision as it's made. A failed check can send the answer back, hold it for a person, or block it.
Boundary: This does not determine whether the risk-management system is legally sufficient.
Article 12
Capability: What went in, how it scored, what was decided and why — all recorded as the decision happens, and chained so it can't be quietly changed later.
Boundary: Illustrative public records are not production certificates or prescribed legal logs.
Article 14
Capability: Mergetic can stop a decision and hand it to a named person before it takes effect — and it records that it did.
Boundary: Workflow design, competence, authority and accountability remain deployment-specific responsibilities.
Article 72
Capability: The records Mergetic keeps can feed your post-market monitoring. They are evidence for it, not the whole monitoring system.
Boundary: The records do not, by themselves, constitute a complete post-market monitoring system, legal plan or compliance determination.
The AI Act's transparency phase is here—and the obligations financial-services boards already carry never waited for it. Supervisors from the Central Bank of Ireland to the PRA continue to expect accountable governance and evidence around AI and model risk.
PRIMARY SOURCES · THEN THE PROOF
The consolidated Regulation and the 2026 amendment are linked below, direct from EUR-Lex. Then run the demo and watch one governed decision produce the kind of decision-time evidence these articles describe.
