No one sees the full process
AI tools are used across teams while a central register of content types, roles and decisions is missing.
Independent initiative · Pilot phase
Make AI-content visible and accountable.
ShowAI translates European transparency rules into a workable publishing process: identify, review, document and publish responsibly.
PUBLISHING WORKFLOWAccountable
Record AI useTool, content type and owner✓
Review the contentHuman review and editorial control✓
Determine disclosureLabel, notice or another measure✓
Document the decisionVersion, date and rationale✓
No black box. Each step shows who decided, on what basis and what was checked before publication.
Article 50as the legal core
Human reviewas a process step
Content registeras an evidence structure
Not legal adviceclear boundaries
The problem
AI use is growing faster than internal rules.
Organisations increasingly publish text, images, audio and video created with generative AI. Without a fixed process, review, transparency and accountability become inconsistent.
Disclosures are improvised case by case
Employees do not always know when transparency is legally relevant or voluntarily appropriate.
Human control remains invisible
Without consistent records, it is difficult to show who substantively reviewed and approved content.
Why it matters
Weak documentation becomes costly when questions arise.
From 2 August 2026, organisations within the scope of Article 50 must meet the applicable transparency obligations. A clear record helps teams make consistent decisions and explain what happened when clients, auditors or authorities ask.
Harder to demonstrate decisions
Without a consistent record, an organisation may struggle to show why content was disclosed, not disclosed or routed through human review.
More exposure when complaints arise
Natural and legal persons may submit complaints to a market-surveillance authority. Poor records make a response slower and less reliable.
Costly reconstruction
Teams may need to search old files, interview staff, review past publications and obtain external legal support to reconstruct what happened.
Loss of trust
Inconsistent AI disclosures can create client concerns, contractual friction and reputational damage even where no formal penalty is imposed.
Important boundary: Article 50 does not create a universal legal duty to keep a complete log for every item of AI-assisted content. Documentation is valuable because it supports consistent decisions and evidence, especially where an organisation relies on human review and editorial responsibility.
Inventory
Map AI tools, content types, publishing channels and the roles involved.
Process analysis
Examine the real workflow: who creates, reviews, changes, approves and publishes?
Decision matrix
Translate recurring situations into consistent questions, outcomes and escalation points.
Review and evidence structure
Set up a practical register, review procedure and allocation of responsibilities.
Implementation plan
Define concrete actions for adoption, staff guidance and periodic control of the publishing process.
Process
From scattered AI tools to one manageable workflow.
The service begins with the real decisions people must make before publication, not with software.
Intake
Collect information on use cases, audiences, channels and current policies.
Analysis
Trace several real content routes from creation to publication.
Design
Design decision questions, roles, review steps and documentation.
Handover
Deliver templates and concrete implementation actions to the team.
Who it is for
For organisations that professionally manage public-facing content.
Communications agencies
For agencies producing AI-assisted content for multiple clients.
Public-affairs teams
For teams publishing about politics, policy and matters of public interest.
Civil-society organisations
For organisations that value public trust and demonstrable accountability.
Publishers and editorial teams
For smaller publishing organisations that want to document human oversight consistently.
ShowAI Learn
Practical answers to specific Article 50 questions.
The knowledge base separates binding law, official guidance, voluntary measures and practical recommendations — with primary sources and review dates.
First publication
What is Article 50 of the EU AI Act?
A source-based guide to provider and deployer duties, deepfakes, public-interest text, human review and the 2026 application dates.
Read the full guide → CodeIs the Code of Practice mandatory?
Why the Code is voluntary while applicable Article 50 obligations are not.
Read the explanation → WorkflowWhat does human review mean?
How substantive review, authority and editorial responsibility differ.
Read the explanation → Risk & evidenceWhy do transparency records matter?
The legal penalty framework and the practical cost of weak documentation.
Read the explanation →
About ShowAI
European rules translated into everyday publishing decisions.
ShowAI is an independent initiative by Lotta Punt at the intersection of AI, European regulation and public communication. The method is in its pilot phase and is being developed using official sources and practical cases.
Pilot
Bring one publishing process into focus.
The initial pilot is intended for a small organisation that wants to examine and structure its current AI-content process.