How we use AI
We build AI systems for our clients – and we use AI in our own work as well. Whatever is created with the help of generative AI, we label it.
This page explains where generative AI feeds into our work and what the notices on our blog articles and product images mean.
Why we label
Since 2 August 2026 the transparency obligations under Article 50 of the EU AI Act apply. Among other things they require AI-generated or AI-manipulated content to be recognisable as such – for text on matters of public interest just as much as for deceptively realistic image, audio and video content.
The regulation provides an exemption for editorially reviewed content with a named responsible person. We label it anyway – because traceability is not a special case for us, it is the default.
Where we use AI
Generative AI is part of our daily work. In concrete terms, it affects these four areas.
Blog articles
Our weekly AI and tech news, as well as the in-depth technical articles, are written from researched sources with AI support at the drafting stage. Every article is reviewed, corrected and released by the named author before publication.
Images and illustrations
Header and illustration images for our blog articles, as well as the Habicht AI product imagery, are generated with Midjourney. They do not depict real people, places or events. Screenshots, charts based on real data, logos and team photos are excluded.
Software development
In development we use AI assistance for drafts, refactorings and tests. Every line goes through review and testing by our team before it reaches a client project.
Research and summaries
We use language models to work through public sources – release notes, standards, studies. Facts and figures are checked against the original source, which we link in the article.
What the labels mean
You will find these notices on our blog articles and on AI-generated product imagery.
The text was drafted with AI support and editorially reviewed before publication. Responsibility for the content lies with the person named in the article.
The image was created with generative AI. The notice sits directly on the image; on blog articles the caption additionally names the tool used.
Machine-readable labelling
In addition to the visible notice, we mark AI-generated images in the article's structured data. We use the IPTC value “trainedAlgorithmicMedia” as the digitalSourceType – the same standard search engines and image archives evaluate.
Frequently asked questions
Which content on this website is AI-assisted?
The German-language AI and tech news series and the in-depth technical articles from 2026 are drafted with AI support and editorially reviewed; their images are AI-generated. Older articles were written by hand – where AI-generated illustrations were used there, that is labelled too. The Habicht AI product imagery is generated as well. All other pages, meaning services, references, jobs and legal texts, are written by people.
Who is responsible for AI-assisted articles?
The author named in the article. Editorial responsibility always rests with a person, not with the tool. If you spot an error, write to us – we will correct it and note the change.
Is client data used for AI?
No. Confidential client data and project source code never reach public AI services and are not released for training third-party models. In client projects we run AI exclusively on infrastructure that is jointly defined and documented.
On what legal basis do you label?
Article 50 of the AI Act (Regulation (EU) 2024/1689) governs transparency obligations for AI systems; these obligations have applied since 2 August 2026. The regulation provides an exemption for editorially reviewed content with a named responsible person. We label it regardless, because traceability is part of the standard we hold ourselves to.
Questions about AI compliance in your own company?
We build AI systems where transparency, documentation and traceability run along from the start – on-premise or in an EU data centre.
You will find the labelled articles in our blog