This article was collected and archived by Digital Sovereignty Watch from an institutional or public source relevant to digital sovereignty, technology policy, cybersecurity, cloud services, artificial intelligence or European regulation.
AI
Futurium | Apply AI Alliance - AI Transparency Package = Guidelines + Code of Practice
This week, the European Commission published the Guidelines on AI transparency obligations under Article 50 of the AI Act. They’re aimed at providing guidance to providers, deployers and competent authorities in understanding what is in and out of scope, definitions, practical examples and exemptions ahead of the 2 August entry into application.
As AI-generated content and interactive AI systems become more advanced, people need to know when they are:
Exposed to certain AI systems such as emotion recognition and biometric categorisation
The new Guidelines support providers, deployers and competent authorities in understanding how Article 50 applies in practice. The voluntary Code of Practice offers a practical pathway to compliance, including EU icons for labelling AI-generated content.
Transparency is essential to help people make informed decisions and to foster a trustworthy environment for AI innovation in Europe.
You can show your commitment to transparent, trustworthy AI and streamline your pathway to compliance by becoming a signatory to the Code of Practice.
Those involved in creating the article 50 regulation appear to have had zero experience of working with these tools in a market context.
This regulation targets the tool rather than the intent or harm. Bad actors will ignore the labeling requirement, while compliant European startups and small businesses lose a major cost advantage that helped them compete with large companies.
Europe will be the last market for Photoshop. (You don't need a badge if you created the same content with Photoshop.)
Responsibility for the content, not whether AI was used to create it, should be the deciding factor.
Andrea, thank you for sharing. This is a timely and important package.
Article 50 tells users that content is AI-generated. But it does not yet tell them how that content can be verified — what sources it relied on, what reasoning path it followed, and under what conditions it may be unreliable.
We are building exactly that missing layer: a deployable trust infrastructure that ensures every AI output carries traceable evidence chains, auditable decision trails, and explicit confidence boundaries.
If any of the signatories to the Code of Practice are looking for technical pathways to go beyond labeling — to make their AI outputs not just transparent, but truly verifiable — we would welcome the conversation.
As AI-generated content and interactive AI systems become more advanced, people need to know when they are:
Exposed to certain AI systems such as emotion recognition and biometric categorisation
The new Guidelines support providers, deployers and competent authorities in understanding how Article 50 applies in practice. The voluntary Code of Practice offers a practical pathway to compliance, including EU icons for labelling AI-generated content.
Transparency is essential to help people make informed decisions and to foster a trustworthy environment for AI innovation in Europe.
You can show your commitment to transparent, trustworthy AI and streamline your pathway to compliance by becoming a signatory to the Code of Practice.
Those involved in creating the article 50 regulation appear to have had zero experience of working with these tools in a market context.
This regulation targets the tool rather than the intent or harm. Bad actors will ignore the labeling requirement, while compliant European startups and small businesses lose a major cost advantage that helped them compete with large companies.
Europe will be the last market for Photoshop. (You don't need a badge if you created the same content with Photoshop.)
Responsibility for the content, not whether AI was used to create it, should be the deciding factor.
Andrea, thank you for sharing. This is a timely and important package.
Article 50 tells users that content is AI-generated. But it does not yet tell them how that content can be verified — what sources it relied on, what reasoning path it followed, and under what conditions it may be unreliable.
We are building exactly that missing layer: a deployable trust infrastructure that ensures every AI output carries traceable evidence chains, auditable decision trails, and explicit confidence boundaries.
If any of the signatories to the Code of Practice are looking for technical pathways to go beyond labeling — to make their AI outputs not just transparent, but truly verifiable — we would welcome the conversation.