AI and Trademark Law: Can a Brand Created by an AI be Protected by the Law

Introduction

Artificial intelligence (AI) has caused a paradigm shift in various areas, including the field of law. One of these is intellectual property (IP) law, and by extension, trademark law. The rapid growth of AI has managed to redefine many elements of human existence, with the law of intellectual property being a fundamental frontier for technological advancement and legal comprehension.

This detailed study is a deep dive into the intricate relationship between artificial intelligence and trademark law, exploring the legal, ethical, and practical implications of AI-generated brands across different jurisdictions. As businesses shift to leverage AI software to generate brand identities, logos, and advertising campaigns, the question of whether or not these AI-generated assets can meet the criteria for legal protection becomes increasingly pertinent. This question will try to unravel the complex interactions of law, technology, and creativity in an effort to provide clarity on this evolving issue.

The Technological Context of AI-Generated Trademarks

Artificial intelligence has travelled far from a theoretical phenomenon to a realistic technology capable of producing sophisticated, creative works. Machine learning algorithms and generative AI models can now produce logos, brand names, and distinctive signs with greater sophistication than ever before. The AI systems are founded on large pools of data, neural networks, and complicated algorithms that produce unique identifiers that challenge traditional assumptions of authorship and creativity.

Modern AI trademark creation usually consists of some of the following processes:

1. Ingesting and processing existing trademark databases
2. Identifying successful branding patterns
3. Generative algorithms creating new patterns
4. Semantic and linguistic processing for relevance and marketability

The Evolution of AI in Creative Processes

The trajectory of AI development in creative fields has been remarkable. Early AI systems were rule-based and could perform only exact, predetermined tasks. However, current AI systems, particularly those using deep learning and neural networks, can analyse patterns, learn from examples, and generate new outputs that sometimes rival human creativity.

For trademark creation specifically, AI systems like GPT-4, DALL-E, and Midjourney can now:

– Create unique brand names based on industry keywords

– Create visually distinct logos with appropriate colour schemes and typography

– Develop overall brand identities with consistent visual language

– Test consumer responses to various brand elements based on predictive analytics

This technological feat raises fundamental questions regarding the nature of creativity itself. Is creativity uniquely a human attribute, or is it replicable by sophisticated algorithms? The ramifications of how one responds to this question are enormous for trademark law, which has conventionally been premised upon human creative input.

AI and Intellectual Property (IP): Fundamental Principles

AI is able to create new works and marks on its own without the need for human intervention. This makes the legal ownership rights of AI-created marks complex. IP laws have been authored traditionally to protect human ingenuity. However, AI-created trademarks challenge this human-cantered framework and demand reconsideration of existing laws.

Artificially intelligent-generated trademarks refer to trademarks that are generated with the help of AI and without human involvement or with minimal human involvement. They can be brand names, logos, or any other sign of distinction. Whether such marks are eligible for protection under current trademark law or not is a controversial topic since authorship, originality, and ownership are involved.

The Difference Between AI Being Used as a Tool and Being Used as a Creator

A fundamental distinction in this discussion is whether AI is merely a tool in the hand of a human (such as design software) or whether it is itself an independent creator. This makes all the difference in law:

AI as a Tool: Where AI is being employed more as a sophisticated tool at the behest of human users, then the creative work may still belong to the human operator, just as a photographer would own the copyright on photographs taken using a camera.

AI as a Creator: Where AI programs have extensive autonomy in the manner in which they generate outputs with minimal human input, then traditional notions of authorship are called into doubt.

Most legal systems assume the first model tacitly, but technological advancement is progressively pushing towards the second scenario, creating legal tensions that must be resolved.

Legal Frameworks: A Comparative Analysis

UK Trademark Act 1994 and AI-Generated Trademarks

The UK Trademark Act 1994 provides protection for distinctive trademarks with the capability to differentiate goods or services of an undertaking from others. The Act does not, however, explicitly address the case of AI-generated trademarks.[1]

Issues of particular concern are:

– Deciding ‘personhood’ when AI files for trademark

– Deciding intention to use a trademark

– Deciding originality of AI-generated marks

An application for a trademark registration under the Act is to be made by a ‘person’ who claims to be the owner of the trademark. ‘Person’ is not defined in the Act and so must be interpreted. It usually means an individual or legal person such as a company. This has consequences for whether an AI system can or cannot be a ‘person’ under the Act.

Aside from this, the Act also demands that the applicant have an intention to use the trademark. This is a human intent and the same cannot be attributed to an AI system. Therefore, under the current interpretation of the Act, there are questions whether AI-created trademarks can be registered.

Recent Developments in UK Case Law

Recent UK case law has begun to address intellectual property aspects of AI, although not specifically AI-generated trademarks. In copyright proceedings, the UK courts have demanded copyright protection with human authorship. By analogy, this point can be made under trademark law, and it can be stated that AI-only created signs lacking any appreciable human contribution may face problems in registration.[2]

The UK Intellectual Property Office (UKIPO) has launched consultations on AI and IP rights, which points towards an acknowledgment that policy development in this field is necessary. Though there is no clear guidance yet that is unique to AI-generated trademarks, the direction of travel is towards taking a conservative route that maintains human involvement in the creative process.[3]

EU Trademark Law (EUTMR) and AI-Generated Trademarks

The same concerns are also raised by the European Union Trademark Regulation (EUTMR). Though more lenient in some interpretations, the basic requirements are still human-centric:

– Requirement of a natural or legal person as an applicant

– Evidence of distinctive character

– Graphical representability of the trademark

EUTMR safeguards those trademarks that are capable of graphical representation and are distinctive. Nonetheless, similar to the UK Trademark Act, EUTMR also remains silent when it comes to AI-generated trademarks.[4]

Under the EUTMR, an application for a trademark must be filed by a ‘natural or legal person’. This also raises the issue of whether an AI system would be considered a ‘person’ under the EUTMR. Aside from this, the EUTMR also requires that the applicant must have a bona fide intention to use the mark. This intention, similar to under the UK Act, is a human factor and cannot be delegated to an AI system.

The EU’s Regulatory Strategy towards AI

The European Union has been at the forefront of developing comprehensive AI regulations with initiatives like the pending AI Act. Although directed primarily at AI safety and ethical use, these regulations have implications for intellectual property. The EU’s approach emphasizes human oversight and accountability, and this may influence how AI-generated trademarks will be treated going forward.

The European Union Intellectual Property Office (EUIPO) also conducted studies on the impact of AI on IP rights.[5] These studies acknowledge the growing role played by AI in producing potentially protectable content while affirming the need for human involvement in the creative process.

US Trademark Law and AI-Generated Marks

The United States Patent and Trademark Office (USPTO) is controlled by the Lanham Act, which, like its European equivalents, wasn’t written with AI-created content in consideration. Some of the most significant elements in US trademark law include[6]:

– The requirement of “use in commerce” or intent to use

– The requirement that the mark be non-functional and distinctive

– The concept of the trademark as a source identifier

Current USPTO guidance is that AI programs cannot be identified as the inventor or author of IP that can be protected. By extension, this would also imply that fully AI-generated trademarks with insufficient human direction can expect issues being registered in the US.

However, if a human or legal entity uses an AI-created mark in commerce for the identification of goods or services, the mark may be registered as long as ownership is asserted by the human or entity. In this way, use of the mark, and not how it was created, may prove to be the tipping point under US trademark law.[7]

World Intellectual Property Organization (WIPO) and AI-Generated Trademarks

The World Intellectual Property Organization (WIPO) has also been grappling with the issue of AI and IP. Their recent discussions indicate a range of important considerations:

– The need for malleable intellectual property regimes

– A balance between legal protection and innovation

– Finding means of identifying creative works produced by AI

WIPO initiated IP policy and artificial intelligence consultations in 2019, requesting submissions from member states and other interested parties. WIPO has yet to take a definitive stance on the issue of AI-generated trademarks, however.[8]

WIPO’s Ongoing Consultations

WIPO continues to hold international discussions on AI and IP through its Conversation on Intellectual Property and Artificial Intelligence. Several major themes have arisen from these talks:

– The need for international harmonization of approaches to AI-generated IP

– The need to maintain incentives for innovation while adapting to technological change

– The importance of transparency in AI-created works

While these consultations have not yet resulted in binding international treaties, they are shaping the international discussion on how IP systems need to adapt to AI technologies.[9]

Technological and Legal Challenges

Authorship Rights in AI-Created Works

Authorship of AI works has close nexus with the problem of AI-generated trademarks. Authorship of a work typically vests in a human author under most IP laws. In the case of AI-generated works, however, the authorship is unclear.

Some relevant questions include:

– Is the author the AI program itself?

– The creator of the AI who coded it?

– The person who commissioned generation of the trademark?

– The firm owning the AI technology?

The Chain of Creation and Attribution

The following are parties that may be involved in creating an AI-generated trademark:

1. AI Developers: Engineers and programmers that create the AI system
2. Training Data Providers: Parties that gather and supply the data the AI is trained on
3. System Operators: Individuals that install and instruct the AI to generate specific outputs
4. End Users: Those who select, customize, and ultimately use the AI-generated trademarks

Each of these stakeholders might potentially claim some form of authorship or right of ownership, complicating the legal scenario.[10] Some jurisdictions might take into account contributions along the whole chain, while others might take into account only particular links solely.

AI and Human Co-Authorship

Another possible solution to the authorship right issue in AI-generated works is AI and human co-authorship. In this arrangement, both the AI system and the human user would share co-ownership of authorship over the work.[11]

This solution, nevertheless, poses additional questions:

– In what way would co-authorship rights and obligations be divided?

– What are the mechanisms that would ensure fair attribution?

– In what manner would profits be distributed?

Practical Models for Co-Authorship

Different models of handling human-AI collaboration have been proposed:

1. The Tool Model: AI is a highly sophisticated tool, and rights reside with the human user
2. The Employee Model: AI is an employee creating work-for-hire, and rights belong to the employer
3. The Collaboration Model: Rights are divided among the AI creator, the user, and maybe other parties
4. The Public Domain Model: AI-generated works automatically fall into the public domain

The consequences of each model differ with regard to trademark registration and enforcement. The tool model best aligns with current legal frameworks but may be increasingly strained as AI systems develop towards more autonomy.[12]

Technical Verification Challenges

A practical challenge to registering AI-generated trademarks is verification. Trademark offices would need systems to:

– Determine that a mark was AI-generated or created by a human

– Ascertain the extent of human guidance in the process of creation

– Ask whether the extent of human contribution is enough to meet the threshold of protection

Current technical constraints make this verification difficult, which is a source of concern in terms of transparency and accountability in the process of trademark registration.

Ethical and Commercial Implications

Market Impact of AI-Generated Trademarks

The ubiquity of AI-generated trademarks can have significant market implications:

– Trademark Saturation: AI systems can generate and register numerous potentially valuable trademarks, perhaps leading to trademark squatting or a scarcity of available marks

– Market Confusion: If the distinctiveness of AI-generated marks is uncertain, this can lead to more consumer confusion

– Competitive Advantage: Entities with access to advanced AI systems can gain unfair advantages in securing valuable brand identities

These marketplace trends challenge whether trademark systems that exist today, designed for human creativity at a human pace, will be capable of handling the potential flood of AI-generated applications.[13]

Consumer Protection Issues

Trademark law is as much about protecting consumers from confusion as it is about brand owner protection. AI-generated trademarks raise several consumer-centric issues:

– Whether AI-generated marks can effectively indicate source

– The potential for AI systems to generate marks that are subtly similar to existing trademarks, potentially leading to consumer confusion

– Whether to require disclosures of AI use in trademark creation

These considerations can influence how courts and trademark offices evaluate AI-created marks, particularly in likelihood of confusion analyses.[14]

Future Outlook

The future of AI and trademark law will likely involve:

– Continued technological advancement of AI creativity

– Ongoing legal rebalancing to accommodate new paradigms of creativity

– Policymaker, legal, and technologist interdisciplinary coordination

– Creating custom registration pathways for AI-assisted creative works

Potential Legal Reforms

The potential legal reforms below can mitigate challenges of AI-generated trademarks:

1. Modified Personhood: The law can be expanded to give certain AI systems limited legal personhood under IP laws
2. Registration Disclosure Requirements: Applications for trademarks can require disclosure of the use of AI in creation
3. Differentiated Protection Regimes: Gradated levels of protection can be offered based on the extent of human creative input
4. New Sui Generis Rights: New IP rights can be created specifically for AI-generated content

These reforms would require significant legislative change but can better accommodate the realities of AI creation.

Industry Best Practices

As legal systems evolve, industry best practices are emerging to manage today’s uncertainty:

– Making allowance for substantial human guidance and decision-making in AI-created trademark creation

– Maintaining detailed records of the inventive process, including human inputs and decision points

– Conducting thorough clearance searches for AI-created marks, with sensitivity to potential similarity to other marks

– Being transparent with trademark offices about AI involvement in the inventive process

These practices can enable businesses to leverage AI capabilities with minimum legal risks under the current trademark regimes.

Conclusion

AI-generated trademarks are a fascinating area of legal and technological development. The UK Trademark Act 1994 and EU Trademark Law (EUTMR) do not address this issue specifically but, according to contemporary interpretation, seem to leave limited scope for the protection of purely AI-generated trademarks with no significant human input.

As the technology of AI continues to evolve, legal frameworks will need to adapt to provide consistent, flexible regimes that balance innovation and protection. Getting to complete regulation of AI-generated trademarks will take ongoing debate, innovative thinking, and willingness to revisit fundamental definitions of legal concepts to reflect changing technological possibilities.

The most reasonable direction is to recognize the human elements in the AI creative process—whether as developers, prompt engineers, curators, or users—while also considering the unique contribution of AI systems. This hybrid approach keeps intact the human-centred foundations of intellectual property law while leaving space for technological innovation.

In short, while AI can revolutionize the field of IP, it also presents difficult legal and ethical questions that must be resolved with care. As AI is still evolving, it is essential that IP laws evolve with it so that they remain pertinent and effective within the digital economy. The future decade will most likely see significant advancements in this area as cases progress through the courts and legislators respond to the threats and promise of AI-created creative content.

Reference(S)

1. Trade Marks Act 1994
2. Lanham Act 1946, 15 USC § 1051
3. European Union, Regulation (EU) 2017/1001 of the European Parliament and of the Council of 14 June 2017 on the European Union trade mark [2017] OJ L154/1
4. European Commission, ‘White Paper on Artificial Intelligence: A European Approach to Excellence and Trust’ COM(2020) 65 final.
5. UK Intellectual Property Office, ‘Artificial Intelligence and Intellectual Property: Copyright and Patents’ (2021) GOV.UK https://www.gov.uk/government/consultations/artificial-intelligence-and-intellectual-property-copyright-and-patents accessed 6th April 2025
6. European Commission, ‘White Paper on Artificial Intelligence: A European Approach to Excellence and Trust’ COM(2020) 65 final.
7. World Economic Forum, ‘Artificial Intelligence Collides with Patent Law’ (2018) <http://www3.weforum.org/docs/WEF_48540_WP_End_of_Innovation_Protecting_Patent_Law.pdf> accessed 7th April 2025
8. WIPO, ‘WIPO Conversation on Intellectual Property and Artificial Intelligence’ (2023) <https://www.wipo.int/about-ip/en/artificial_intelligence/conversation.html> accessed 6th April 2025
9. WIPO, ‘WIPO Technology Trends 2019: Artificial Intelligence’ (2019) <https://www.wipo.int/edocs/pubdocs/en/wipo_pub_1055.pdf> accessed 7th April 2025

[1] Trade Marks Act 1994

[2] UK Intellectual Property Office, ‘Artificial Intelligence and Intellectual Property: Copyright and Patents’ (2021) GOV.UK https://www.gov.uk/government/consultations/artificial-intelligence-and-intellectual-property-copyright-and-patents accessed 6th April 2025

[3] UK Intellectual Property Office, ‘Artificial Intelligence and Intellectual Property: Copyright and Patents’ (2021) GOV.UK https://www.gov.uk/government/consultations/artificial-intelligence-and-intellectual-property-copyright-and-patents accessed 6th April 2025

[4] Regulation (EU) 2017/1001 of the European Parliament and of the Council of 14 June 2017 on the European Union trade mark [2017] OJ L154/1.

[5] European Commission, ‘White Paper on Artificial Intelligence: A European Approach to Excellence and Trust’ COM(2020) 65 final.

[6] Lanham Act 1946, 15 USC § 1051

[7] United States Patent and Trademark Office, ‘Public Views on Artificial Intelligence and Intellectual Property Policy’ (2020) <https://www.uspto.gov/sites/default/files/documents/USPTO_AI-Report_2020-10-07.pdf> accessed 6th April 2025

[8] WIPO, ‘WIPO Conversation on Intellectual Property and Artificial Intelligence’ (2023) <https://www.wipo.int/about-ip/en/artificial_intelligence/conversation.html> accessed 6th April 2025

[9] WIPO, ‘WIPO Conversation on Intellectual Property and Artificial Intelligence’ (2023) <https://www.wipo.int/about-ip/en/artificial_intelligence/conversation.html> accessed 6th April 2025

[10] World Economic Forum, ‘Artificial Intelligence Collides with Patent Law’ (2018) <http://www3.weforum.org/docs/WEF_48540_WP_End_of_Innovation_Protecting_Patent_Law.pdf> accessed 7th April 2025

[11] World Economic Forum, ‘Artificial Intelligence Collides with Patent Law’ (2018) <http://www3.weforum.org/docs/WEF_48540_WP_End_of_Innovation_Protecting_Patent_Law.pdf> accessed 7th April 2025

[12] WIPO, ‘WIPO Technology Trends 2019: Artificial Intelligence’ (2019) <https://www.wipo.int/edocs/pubdocs/en/wipo_pub_1055.pdf> accessed 7th April 2025

[13] World Economic Forum, ‘Artificial Intelligence Collides with Patent Law’ (2018) <http://www3.weforum.org/docs/WEF_48540_WP_End_of_Innovation_Protecting_Patent_Law.pdf> accessed 7th April 2025

[14] World Economic Forum, ‘Artificial Intelligence Collides with Patent Law’ (2018) <http://www3.weforum.org/docs/WEF_48540_WP_End_of_Innovation_Protecting_Patent_Law.pdf> accessed 7th April 2025