本文来自微信公众号： 金杜研究 ，作者：瞿淼，原文标题：《China in a Changing World Series | New Challenges in Brand Protection in the AI Environment》
Introduction
Artificial intelligence is profoundly reshaping the landscape of brand protection.This "double-edged sword" presents unprecedented challenges:on one hand,the methods and channels for brand exposure are undergoing profound changes,while the most fundamental role of a brand is to indicate its source,leading to essential shifts in both brand marketing paths and high-risk infringement scenarios;on the other hand,AI tools have drastically altered the means and methods of counterfeiting and confusion,necessitating adjustments to traditional rights protection and management approaches.
01
New Brand Issues in the AI Environment
1.Trademark Infringement Risks of AIGC
The focus on AIGC has always been on copyright,which is a decisive battle of interests between copyright law and AI technology companies.However,another non-negligible scenario is trademark infringement in AI-generated content caused by AI training.[1]If a company uses AI tools to design a trademark,it may "unconsciously" imitate the composition,colors,or visual elements of an existing trademark,generating a mark highly similar to another's registered trademark.The Getty Images v.Stability AI case is a landmark case in this field.Therefore,after extensive use of AIGC,the risk of brand infringement is as severe as that of copyright infringement.
2.The Risk of Trademark Genericide is Significantly Amplified
Article 49 of China's Trademark Law stipulates:"Where a registered trademark has become the generic name of the goods in respect of which it is registered,or where the trademark registrant has not used the registered trademark for three consecutive years without proper reasons,any entity or individual may apply to the Trademark Office for cancellation of the registered trademark."
In the AI environment,AI chatbots may use brand names as generic terms when answering questions(e.g.,"search using Baidu").Such usage is highly likely for original technology creators(such as OPEN CLAW)or absolute market share leaders in a field(such as Google,Baidu),leading to generic use of the brand name by users.If such usage occurs frequently and involves a large population,it may accelerate the risk of the brand becoming generic.Moreover,as the behaviour of AI users constitutes individual actions,it is often difficult to monitor through conventional channels.
3.The Foundation of Traditional Brand Premium and the Attention Economy is Being Subverted
The core function of traditional brands lies in indicating source and reducing consumer decision-making costs.Consumers face a significant information gap when making purchasing decisions.Even in the internet era,they encounter information gaps and "information cocoons" created by platform algorithm recommendations.However,they can make purchasing decisions based on brand trust and are willing to pay a brand premium for it.
In the AI era,consumers,empowered by the infinite information processing capabilities of AI tools,can not only deeply analyse their characteristics and preferences based on comprehensive personal information but also effortlessly traverse vast amounts of product information across platforms and conduct data comparisons on various aspects.The traditional attention economy model of e-commerce platforms(monopolising attention entry points,paid search rankings,algorithmic recommendations,etc.)will be severely impacted.This AI-assisted consumer decision-making mechanism will significantly enhance consumers'ability to make informed decisions and achieve precise person-product matching.The brand advantages and premium mechanisms established through reliance on traditional marketing investments will be challenged.
02
New Dilemmas Faced in Brand Protection and Enforcement
1.Changes in Infringement Methods and Means
The emergence of new AI tools also means they will be used as means and instruments for infringement,giving rise to new types of infringing acts.For example:(1)AIGC may produce and disseminate incorrect brand information due to "hallucinations," which is difficult to monitor and has a long correction cycle;(2)AI tools may drastically reduce the cost of counterfeiting and significantly increase the efficiency of dissemination.For instance,black market operations have already used AI to generate realistic images of "damaged goods" or "stains" in bulk to conduct large-scale extortion by exploiting platform "refund-only" rules;(3)The boundary between physical and virtual goods may become increasingly easy to breach;(4)The traditional territorial nature of trademark protection may be completely overturned in the AI environment.In response to these new phenomena and methods,traditional rights protection tools and systems are no longer effective and face reconstruction.
2.The Legal Liability System in the AI Environment is Not Yet Established
As discussed earlier,the essential role and dissemination channels of brands have changed significantly in the AI environment.Model operators and agent platforms will likely have to assume responsibilities similar to those of internet platform operators,becoming the primary subjects of legal regulation and key hubs for rights protection.Currently,as AI business models are still in the early stages of development,the legislative system will likely be shaped again through judicial practice.Operators may need to rely on the existing judicial system,using litigation and other means,to clarify the legal legitimacy of various actions and delineate the boundaries and principles of legal liability for different entities.
3.The Criteria for Determining Similarity of Marks and Classes in Trademark Infringement May Change
In the AI era,due to further changes in dissemination methods,the standards and principles used in trademark infringement determinations may change.For example,(1)The criteria for determining similarity of trademarks have traditionally been based on standards formulated for human observation with the naked eye under normal conditions.With the advent of AI tools,if most consumers'purchasing behaviours in the future are realised through AI agents or other tools,should the identification capabilities of AI tools(non-human)be considered in the determination standards?(2)The traditional International Classification of Goods and Services for the Purposes of the Registration of Marks under the Nice Agreement(known in China as the "Classification of Similar Goods and Services")will also face significant challenges.This classification has already been stretched thin in coping with the internet era.For instance,Class 9(computer software)and Class 42(technical services)may be strongly related or even similar to many other classes such as Class 35(advertising,sales promotion)and Class 41(entertainment,content production,etc.).Furthermore,the definition of the "relevant public" may become blurred due to shared underlying AI models and dissemination environments.
03
Response Strategies for Brand Protection
1.Brand Repositioning and Portfolio Strategy
As mentioned above,brand portfolio strategy regarding classes needs to fully consider the relevance of Class 42 to other classes in the AI environment,focusing on the usage needs within that context.Additionally,geographically,the cross-regional characteristics brought by AI tools must be fully considered,aiming for early international portfolio development.Furthermore,to adapt to the working methods of AI tools,more precise brand profiling and positioning targeting consumer characteristics are necessary to facilitate AI recognition and matching.
2.Reconstructing Brand Usage and Marketing Methods
Being recognised and acknowledged by AI will become the core task of brand building in the AI era.Brand content promotion needs to align with AI's crawling and recognition habits:(1)Emphasize the construction of authoritative sources,releasing structured data through key sources such as brand official websites,government databases,and industry-authoritative media to facilitate AI crawling;(2)Construct a knowledge graph about the brand,comprehensively building and linking information such as the brand's historical evolution,corresponding products and their performance,cultural connotations,and the subject entity;(3)Establish authorized cooperation with AI platforms:require mainstream AI tool developers to block or provide warnings for brand keywords and visual elements,and explore an "authorized training" model.
3.Transformation of the Brand Protection System
The AI era necessitates the establishment of a completely new brand protection system,including:
(1)Using technology to counter technology:Employing advanced AI technologies and tools for infringement monitoring.
Pay high attention to issues of brand dilution and genericide.Utilise professional legal teams to conduct high-quality assessments of monitoring information and take appropriate measures.Closely monitor emerging forms and methods of infringement.
(2)Leading enterprises should actively protect their rights to gain rule-making discourse.
In any new field,the construction of a legal system creates opportunities for rule discussion and formulation through the practical actions of participants.Once the rules in a domain have been perfected and established by others first,it becomes significantly more challenging for latecomers to change them.In the AI era,operators of foundational AI models and agent platforms may become key entities.The delineation of their liability boundaries needs to be gradually clarified through various collisions among various industry stakeholders.
(3)Increased emphasis on the cultivation and recognition of well-known trademarks.
From a brand perspective,the cultivation and recognition of well-known trademarks become even more crucial in the AI era.Due to the cross-regional,cross-industry,and cross-platform super information collection,analysis,and processing capabilities enabled by shared model foundations and AI tools,achieving well-known trademark status first can realize higher brand value within the AI environment and secure a higher level of protection.
Conclusion
With the advent of the AI era,we are also ushering in a brand-new era for brand protection work.Enterprises'brand protection efforts must be built upon a profound understanding of the underlying AI technology,shifting from "passive defence" to "proactive planning." Simultaneously,brand protection work,from conceptual awareness and institutional processes to substantive legal issues,will face exciting reshaping.
脚注：
[1]In January 2023,Getty Images,the world's largest stock photo agency,filed simultaneous lawsuits against AI company Stability AI in the High Court of Justice in the UK and the U.S.District Court for the District of Delaware.The lawsuits alleged that Stability AI had copied over 12 million Getty Images watermarked photos without authorization to train its Stable Diffusion model.In November 2025,the UK High Court ruled that the model's weights did not constitute copyright infringement;however,the distorted Getty watermark symbols generated by early versions of Stable Diffusion did constitute trademark infringement.