AI-generated content: Who owns the rights?

Copyright for AI content: Current legal situation and open questions

The nuuse of artificialartificial intelligence (AI) for thecreation of textstexts, images, videos and other creative content weraises many legallegal questions. One of the importanthe most important thing isWho actually ownshe actually owns thecopyright and managementrights to AI-generated inhold? Can I as a user of an AI freely over the results or door do I have to restrictionspay attention? Here a Overview of the current legalsituation in Germanyd.

According to the German Copyright Act (UrhG) only “personal gecreations” of a person are protected by copyright protected. The means: Pure KI-generated contentscontent does not enjoyprotection, since the AI itself is not as the author ofis seen. Also the user who only requestsrequests to the KI is not is not considered as the creator of the content.The lack ofprotection has advantages and disadvantagesdisadvantages: On the one handon the one hand use of KI images you do not name the author or request permission forragen. On the other hand but also exclusive management rightsrights – WApplicants couldcould use the sameuse the same contentzen. This can be problematicif you use the AI-generated inhold commercially usedwould like to no controlabout it haswho still needs the contents used. In additionis the dangerahr that AI-generated inhold against bestExisting copyrightsthe AI with copyright protectedprotected material we traind. The more similar the AI output one existing plant is, the wamore likely is a copyright infringementing.

As a user you shouldshould thereforecheck whether with the use ofthe use of AI contentrights of thirdrights of third partiesrights. Overall, theefindet the copyright due to the emergence of AI in a spfield. EOn the one hand, it should creators and their creativecreations, on theon the other hand, it must innovations too restrictive. It It remains to be seen whether the legislator will in the futureand clearer rulesfor the handling KI-generated inhold. Until there is the use of suchof such contentcaution is advised.

Human contribution as a prerequisite for authorship

Authorship of AI-generated content can only be considered if a human has had a significant influence on the creation of the work. The greater the human contribution, the more likely a copyright can arise. Examples would be the subsequent editing of AI images or the merging of AI texts into a new work. Where exactly the border runs, however, has not yet been conclusively clarified.

To date, there have only been a few court rulings on this topic, so many questions remain unanswered. Different approaches are emerging, particularly at international level. While a court in China recently ruled that AI-generated images can also be protected by copyright, courts in the USA and Europe have so far taken a more restrictive stance. For companies and creative professionals operating internationally, it is therefore important to keep an eye on the legal situation in different countries.

In Germany, the decisive factor is likely to be whether the human contribution reaches the required “level of creation”, i.e. whether it has a certain originality and individuality. This is always a question of the individual case and depends on the type and extent of human involvement. Pure technical support activities such as the selection of training data or the configuration of the AI will probably not be sufficient. It remains to be seen how courts will judge these issues in the future.

Another interesting aspect is the question of whether the use of AI in the creation of works changes the requirements for the level of creativity. Some experts argue that the support of AI would have to raise the bar for human creativity in order to establish authorship. Others see this as an unjustified disadvantage for AI-supported works and argue for a technology-neutral approach. The distinction between copyrights and ancillary copyrights could also play a role in AI-generated content. While copyright is based on personal intellectual creations, ancillary copyrights are linked to investments and entrepreneurial achievements. It is conceivable that special ancillary copyrights could be created for AI-generated content in order to protect the investments behind it, even if the requirements for copyright are not met. Finally, the question arises as to the ownership of copyright and exploitation rights if several people and AI systems were involved in the creation of a work. Similar principles could apply here as for joint works or associated works. The decisive factor is likely to be the creative contribution made by the individual participants and whether a division of rights is possible and reasonable.

Overall, it is clear that the development of AI poses major challenges for copyright law. The previous categories and demarcation criteria do not always fit the new technical possibilities. A careful analysis and possibly also a cautious adaptation of the legal framework is therefore required in order to find an appropriate balance between the interests of all parties involved.

Risk of copyright infringements through AI content

Since KI systems with copyrighted worksprotected worksare trained, there isthere is a fundamental danger that generated contentcontent infringescopyrightslast. The more similar the Output of an existing plant is, the moremore likely is an injury. Groovetors should thereforeI content before publication of thecheck. Pure zual similarities arebut rather unprobably. Problcan become when the AI characteristic elements of a copyrightedwork withoutwithout that this is instruction of the userlooks. Here arestands theof an unconsciousunknowing copyright infringementtion.

On the other hand it should be borne in mind that AI systems with a large number of worksare trained andtherefore rarely exexact copies perduce. Often newindependent contentscontent that from theworkseiden. Nevertheless Caution is advisedoten, especially when you want to use AI contentwant to use AI content commercially.tion. A careful tests examination for possiblecopyright infringements is in any case advisable. If necessary it may be usefulto contractually agreeto ensure that the provider of the AI provider has all the necessaryrights to the trainings data.tion. Also a Exemption from liability for the case of unintentional infringements may come into considerationcome into consideration.Anotherproblem isis that the exact traing data from AI systems are often are often not disclosed. For users It is therefore difficultto understand whichwhich works may have been the generationhave been includedd. Also the funcof the AI itself is often a “black box”, soso that it is not alwaysalways clear according to which criteria contentcontent is generated.

This complicates the check for copyrightcopyright infringementsfurther. In additionis that many KI systems are based on publicly accessibleaccessible data on the Internetare trained without that the rightsholders have have consented to this. This is where the question of whether this is consideredlawful use for the purpose of text and data mining can be considered or whether the rights of the authors losebe used. The The legal situation is on this point not yetfinalizedclarified. Some expare calling foran adaptation of the copyright law in order to the use of protected worksworks for training of AI more clearly regulate. Conceivable wWould legal exceptions forlegal exceptions for certain nucertain uses or the introduction of licenses that remuneration of the rights holders vsee. Also a labeling requirement for AI-generated contentcontent is beingtion to achieve more transparency create more transparency.

Ultimately but it is also responsibility of the users to handle AI contentand to minimizepossible risks in the to keep an eye on. DIn addition to the Testing for originalviolations also the question of whether the generated inhold ethicalretable and inhare correct in the oldd. Especially with sensitive themen or the Creation of Fake news beis the dangerahr that AI systems are problematic contents perduce the uncontrolledwide. Here is a criticalreflection andreview of the results isessential.All in allit is evident that the use of AI for the creation of content not not only legal, butbut also ethicalethical and socialsocial issuesraises questions. It requires a broad discourseand a carefulof the variousinterests in order to a responsible responsible use of this technology find. Thereby must respect both the rights of thetors as well as the innovation potential and the benefits for the general publicare taken into account.

But beware: contractual pitfalls when using AI services

Anyone using AI systems to create content should not only keep an eye on the copyright framework, but also carefully check the contractual agreements with the provider. There are a number of pitfalls lurking here that can restrict the use of the generated content.

It has not yet been conclusively clarified whether and under what conditions providers of AI services may limit the use of the content created in their general terms and conditions. Some clauses stipulate, for example, that users may only use the content for private purposes or may not exploit it commercially unless they have taken out a paid premium subscription. Whether such restrictions are legally permissible depends on a case-by-case assessment. However, users should be aware that by agreeing to the GTC, they may be relinquishing rights that copyright law would actually grant them. Similar problems can arise if you contribute AI-generated content to projects yourself as a contractor or licensee.

As I have already explained in previous blog posts, many contracts contain clauses that grant the client or licensor extensive rights of use to the content created. If you are not careful here, you may be transferring rights that you don’t even have because the content is purely AI-generated. In the event of a dispute, this can lead to liability risks, and agreements that exclude or restrict the use of AI systems in order fulfillment are not uncommon. Particularly in creative industries such as advertising or journalism, some clients insist on a purely human performance. Anyone who secretly resorts to AI here risks not only legal consequences, but also reputational damage. To avoid such pitfalls, it is advisable to check the contractual framework conditions carefully and, if necessary, renegotiate them before using AI systems. Open communication with clients and contractual partners about the use of AI can also help to avoid misunderstandings and conflicts.

Ultimately, everyone involved benefits from a legally compliant and transparent approach to this pioneering technology.

Conclusion: Legal gray areas require vigilance and expert advice

The rapid development of AI systems poses major challenges for copyright law. The existing regulations do not always fit the new technical possibilities, so many questions remain unanswered. This leads to considerable uncertainty for all parties involved, but one thing is certain: purely AI-generated content does not enjoy copyright protection in Germany. Users therefore have a great deal of freedom, but also have to reckon with competition. You can establish copyrights through your own creative contributions, but the boundaries are fluid. In addition, there is always the risk that AI content may infringe third-party copyrights – often without the user’s knowledge.

The situation is exacerbated by unclear contractual regulations. Many providers of AI services try to restrict the use of the generated content, which can lead to conflicts with copyright law. There are also pitfalls lurking in order and license agreements if AI content is introduced carelessly, so great vigilance is required to avoid legal problems. Companies and creative professionals should not only familiarize themselves with the copyright framework, but also carefully review their contracts. It is advisable to seek expert advice at an early stage. Especially in complex or international matters, the expertise of a specialized lawyer can be decisive in minimizing liability risks.

In the long term, there is no way around adapting copyright law to the new circumstances. For example, special ancillary copyrights for AI-generated content or clearer liability rules would be conceivable. Legislators are called upon to facilitate innovation and at the same time create a fair balance of interests. Courts will also be dealing with AI-related disputes and setting precedents more frequently in the future, and as a lawyer I see it as my task to advise clients competently and proactively in this dynamic environment. Because only those who are aware of the opportunities and risks of AI and are on the safe side legally can exploit the full potential of this fascinating technology. I am convinced that with the right strategy and expert support, companies and creative professionals can continue to benefit from AI in the future – in a legally compliant and innovative way.

Author: Marian Härtel

Marian Härtel ist Rechtsanwalt und Fachanwalt für IT-Recht mit einer über 25-jährigen Erfahrung als Unternehmer und Berater in den Bereichen Games, E-Sport, Blockchain, SaaS und Künstliche Intelligenz. Seine Beratungsschwerpunkte umfassen neben dem IT-Recht insbesondere das Urheberrecht, Medienrecht sowie Wettbewerbsrecht. Er betreut schwerpunktmäßig Start-ups, Agenturen und Influencer, die er in strategischen Fragen, komplexen Vertragsangelegenheiten sowie bei Investitionsprojekten begleitet. Dabei zeichnet sich seine Beratung durch einen interdisziplinären Ansatz aus, der juristische Expertise und langjährige unternehmerische Erfahrung miteinander verbindet. Ziel seiner Tätigkeit ist stets, Mandanten praxisorientierte Lösungen anzubieten und rechtlich fundierte Unterstützung bei der Umsetzung innovativer Geschäftsmodelle zu gewährleisten.