EU AI Office issues transparency guidelines and opens Code sign-up

The EU AI Office has published detailed Guidelines on Transparency of AI‑Generated Content and opened sign-up for a voluntary Code of Practice aimed at helping providers and deployers meet Article 50 obligations under the AI Act.

On 22 July 2026 the AI Act Service Desk posted the guidance and a sign-up mechanism, signalling a concerted push by the Commission to operationalise transparency duties ahead of Article 50 taking effect. The documents frame the Code as a voluntary, industry-facing route to demonstrate compliance and set an initial signatory window for public listing of early adopters.

The Guidelines explain the Commission's interpretation of Article 50 and outline the kinds of measures the regulator expects from organisations that generate or disseminate synthetic content. The AI Office positions the Code as a practical route for vendors and deployers to commit publicly to transparency steps such as labelling AI‑generated material, providing provenance metadata where feasible, and adopting internal processes to assess and record how content is produced and distributed.

By opening a short window for inclusion on an "initial list of signatories," the Commission has created an immediate decision point for companies that build or use generative models. Signing early would place a vendor or employer on a visible roster of organisations claiming voluntary adherence, while remaining outside that list could leave firms relying solely on future compliance strategies once Article 50 is enforceable.

The move deliberately targets both upstream suppliers of generative models and downstream deployers — a cohort that includes recruitment platforms, HR tech vendors and employers using AI to create job ads, interview scripts, or employee communications. For HR teams, the sign-up choice is practical: joining the Code could be presented to stakeholders as an interim compliance step, but it also obliges firms to publish how they handle disclosure and provenance for content used with or sent to staff and candidates.

The publication sits alongside a wider regulatory push in the EU to put guardrails around generative AI. Regulators have signalled that transparency obligations will be a core enforcement focus, and the Commission has used the Code mechanism elsewhere as a way to accelerate uptake of best practice while full statutory systems and enforcement plans are finalised.

What the Commission has not detailed will matter to many employers and vendors. The documents do not set out binding audit or verification processes for signatories, nor do they explain whether early signatories will receive any regulatory forbearance or clear shield from penalties once Article 50 is enforced. The criteria for inclusion on the initial list are procedural but leave room for interpretation, and the guidance does not prescribe technical standards for watermarking or metadata formats — instead it points to recommended approaches and cooperative development.

For HR leaders, the sign-up window compresses a compliance timeline: organisations must weigh the reputational and practical benefits of early public commitment against the lack of quantified enforcement protections or independent certification. The next months are likely to be shaped by how quickly major vendors and platform providers declare their positions — their choices will affect downstream employers' ability to rely on provider-led transparency when generating content for recruitment, performance communications or employee learning materials.

The Code's early signatories and any subsequent enforcement guidance will set the tone for how transparent AI‑generated content becomes in everyday workplace interactions. Employers that want to stay ahead will need to map where generative outputs touch hiring and internal communications, decide whether to join the Code or adopt equivalent public commitments, and be ready to adapt disclosure practices as the Commission finalises enforcement details.