Public affairs / EU policy

AI transparency is now an operational communications issue.

Who approved the AI generated content? Does its use need to be disclosed? Can the team show how it was checked? EU transparency guidance brings these questions into everyday communications work.

One9Six Editorial2 min read
Source development: 06/08/2026
Sunlight through transparent partitions in a contemporary institutional chamber.

The development

The European Commission’s guidance, updated on 6 August 2026, states that Article 50 transparency obligations apply from 2 August 2026. It distinguishes providers from deployers and addresses specific AI uses, including interactive systems and synthetic content, with scope conditions and exceptions. European Commission • updated 6 August 2026

The accompanying code of practice is voluntary. That does not make the underlying applicable obligations voluntary. The correct treatment depends on the use case and the organisation’s role. European Commission • AI content transparency code

One9Six / Working notes

Prepare an AI content register

Transparency decisions depend on the use case, the content and the organisation’s role. An operational register helps communications teams identify material that needs specialist review. It is a planning tool, not a legal determination of which obligations apply.

Implementation reference
FocusRecordPractical check
InventorySystem, owner, audience, content type and publication channelInclude generated images, audio and video as well as text
ReviewApplicable role, scenario and current official guidanceRoute uncertain cases to legal and governance owners
ImplementationApproved disclosure, placement and retained evidenceCheck that labels survive distribution and platform formatting

Apply this to your next project

  • Do not assume human editing removes every transparency obligation.
  • Record the guidance version used for each assessment.
  • Give correction and escalation procedures the same owner as publication approval.

One9Six perspective

Public affairs, communications, procurement and legal teams need a shared inventory of AI uses. A chatbot, a synthetic spokesperson and an AI assisted draft create different questions. One blanket label or a supplier’s generic assurance cannot replace that classification.

We recommend that you attach an owner, source record, review decision and disclosure assessment to each workflow. Preserve the approved version and record who has authority to publish or change the system. Obtain legal review for the applicable requirements and exceptions. This briefing is an operational perspective, not a legal opinion.

The next decision

Bring three real AI workflows to a joint review. Identify what is already public, which supplier controls the relevant features and what evidence of review exists. Prioritise workflows that communicate directly with customers or address matters of public interest.

Sources & context

Primary sources checked on 4 October 2026. Recommendations and illustrative examples are One9Six analysis. Product availability and policy guidance may change.

European Commission • updated 6 August 2026

European Commission • AI content transparency code

Apply the method

From AI pilot to accountable marketing workflow

Put the thinking to work.

Explore practical resources

What comes next
for your business?

↗