Imagine a journalist writes an article about a new research finding. A second journalist reads it, researches the original study, and writes their own piece — in their own words, with their own structure. The second article doesn’t infringe the copyright of the first. Why? Because facts aren’t copyrightable. Only the specific creative expression is protected.
Legal distance describes the same mechanism for AI-generated content. When an AI system reads source content, extracts the facts, and writes an entirely new original text without reproducing the source phrasing, it creates a legally independent work. No source copyright is infringed. For digital signage operators, that’s not a theoretical distinction — it’s the difference between a costly licensing contract and a structurally clean operation.
The copyright problem in signage: starting point
Digital signage needs content. Operators looking for current, relevant material have historically turned to three sources:
- Self-created content — demanding, staff-intensive, rarely updated daily
- Licensed news content (wire services) — expensive, contractual, not venue-adapted
- Unlicensed reproductions — cheap, but legally risky to outright illegal
Option 3 is more common than acknowledged. Operators display RSS feeds directly, copy text from websites, or use images without licences. The risk: cease-and-desist letters, damages claims, potentially injunctive relief.
Option 2 is legally clean but expensive. As documented in News feed licensing cost for signage, the all-in cost for a mid-size network can reach £45,000–65,000 over three years.
AI-generated content with verified legal distance is the structurally superior alternative.
What legal distance means in legal terms
Copyright protects the concrete expression of a work — specific word choices, sentence structure, editorial treatment. It does not protect the underlying facts, the events reported, or the news as such.
When an AI system:
- Reads and analyses a source entry
- Extracts the relevant facts and core substance
- Creates an independent text in its own phrasing
- Reproduces no sentence, phrase, or characteristic formulation from the source
…it creates a legally independent original work. The separation is complete.
This isn’t an AI-specific concept. It’s the same mechanism journalists use daily when they cover the same events as colleagues without plagiarising one another.
Three conditions for genuine legal distance
Not every AI application achieves this separation. Three conditions have to be met:
1. No copy-paste, including partial
The system must not reproduce literal text passages from the source — not even individual characteristic phrases. “Slightly rephrased” doesn’t qualify. The output has to be a linguistically independent work.
2. Transformation, not summarisation
A machine-generated summary that shortens or rearranges sentences usually doesn’t meet the requirement. True transformation means new structure, new perspective, new phrasing — with the same facts.
3. Source documentation
To stand up in a dispute, the system has to be able to document which sources were used for which output. That’s the evidence the generated text is an independent work.
Atlas meets all three conditions through its 15-stage pipeline: Stage 7 generates from extracted facts under a system prompt that forbids referencing or paraphrasing the source; Stage 8 runs a deterministic n-gram overlap check against the full source chain and blocks publication at threshold 0.15; Stage 13 writes a SourceTracking record linking every published item to its source, raw item, facts, cluster, and Knowledge Object. The chain reads in both directions, so the audit is operational rather than just documentary.
What legal distance does not mean
A few common misconceptions:
“AI-generated content is copyright-free.” Wrong. AI-generated content has copyright attributes — who the rights holder is (human, organisation, or potentially neither) varies by jurisdiction and is an evolving area of law. The concept means no third-party copyright is infringed, not that no copyright exists.
“Legal distance applies automatically to any AI system.” Wrong. A simple LLM prompt that “summarises” an article does not necessarily meet the conditions. It depends on the specific technical implementation, including the gates between generation and publication.
“With legal distance I can display any content without further restriction.” Wrong. This principle resolves the text copyright issue. Other legal frameworks apply independently: competition law, media law, sector-specific regulation (healthcare, financial services), image licensing.
Legal distance and GDPR: the intersection
The former is a copyright concept. Data protection is a separate dimension. For digital signage screens that don’t process visitor personal data, GDPR relevance is limited — but not zero.
If the AI content engine runs on servers outside the EEA, third-country transfer rules may apply. If the engine builds usage profiles from screen behaviour, consent obligations kick in. An on-premise or EU-hosted solution avoids this complexity by default.
More on the privacy frame: GDPR-compliant AI and Data sovereignty.
The practical value for operators
For a signage operator, legal distance means:
No licence contract required. No annual wire-service fees for content running on your screens.
No infringement risk. As long as the engine genuinely implements legal distance (text generation, n-gram gate, audit log), there’s no exposure to third-party copyright claims on the text.
Scalability without licence tiers. When your network grows from 10 to 100 screens, no proportionally higher licensing cost arises.
Multilingual at no additional cost. If the engine generates in 15 languages, no translation licences are needed.
Auditability. With a production engine, you have source references for every displayed text — your evidence in any dispute, your traceability for compliance.
Practical example: from source to screen
An RSS feed from a health news source contains:
“Scientists at the University of Cologne have demonstrated in a new study that regular walking of 30 minutes per day reduces the risk of heart attack by 22 percent. The study, published in the European Heart Journal, is based on data from 15,000 participants over five years.”
An AI system with legal distance would generate the following for a hotel wellness screen:
“Health tip: just 30 minutes of walking each day can significantly reduce your heart-attack risk — according to a major European long-term study. A stroll through the neighbourhood? Your heart will thank you.”
Same facts. Independent text. No word-for-word reproduction. Adapted for a hotel wellness context. Separation maintained.
Questions to ask when evaluating solutions
When comparing AI content solutions for signage, the legal-distance questions worth asking:
- Does the system actively check for literal text reproduction (and at what threshold)?
- Is there technical documentation of the source-to-output relationship?
- Has the provider obtained legal analysis of their legal-distance claims?
- Is there an audit log for every deployed text linking back to source?
- How are images handled? (The concept applies to text; image licensing is a separate matter.)
How high is your current legal-distance risk?
1. Do you currently use text from news wires or licensed sources on screens?
2. Do you have image licences documented for all screen content?
3. Is screen content reviewed legally on a regular basis?
Screens that keep themselves fresh contboxx delivers licensed news and AI-generated content to your displays, fully formatted — multilingual, automatic, no editorial effort.
Frequently asked questions
What does legal distance mean for AI-generated content?
Legal distance is the copyright separation between source content and AI-generated output. When no source text is reproduced, no creative originality of the source is taken over, and an independent original work is created, the generated work doesn’t infringe the source’s copyright. That’s more than paraphrasing — it’s a structural separation enforced at the data level, not negotiated per item.
Does legal distance apply to all AI systems?
No. It depends critically on the technical implementation. Systems that merely paraphrase or summarise without genuine transformation at the fact level often don’t meet the conditions. The difference between paraphrase and true original text is legally substantive — and liability-relevant for the operator displaying the content. Ask vendors about specific gates (n-gram check, source anchoring, audit log), not just claims.
Do I still need image licences despite legal distance?
Yes. Legal distance concerns text copyright. Images are governed by separate copyright rules, and AI-generated text doesn’t automatically extend to AI-generated visuals. A production content engine should use AI-generated imagery, clearly licensable stock images, or proprietary photography — and document image provenance separately from text provenance.
How does legal distance differ from a news-agency licence?
A news-agency licence permits the use of specific copyrighted texts for a fee — per source, per use, often per screen. Legal distance eliminates the copyright issue structurally by ensuring no third-party work is used, only an independent original work is created. Instead of accepting licensing costs, you avoid the licensing requirement entirely.
Could courts still find AI-generated texts to infringe?
Depends on the specific implementation and the jurisdiction’s case law. With rigorous technical implementation (no literal reproduction, genuine fact-level transformation, complete source documentation, no reproductive proximity to the original), the risk under current German and EU legal frameworks is very low. For regulated industries, specific legal advice on individual high-risk content categories remains advisable.