Artificial Intelligence and Intellectual Property Legal Frameworks in the Asia-Pacific Region

Globally, governments are grappling with the emergence of artificial intelligence (“AI”). AI technologies introduce exciting new opportunities but also bring challenges for regulators and companies across all industries. In the Asia-Pacific (“APAC”) region, there is no exception. APAC governments are adapting to AI and finding ways to encourage and regulate AI development through existing intellectual property (“IP”) regimes and new legal frameworks.

AI technologies aim to simulate human intelligence through developing smart machines capable of performing tasks that require human intelligence. The expanding market for AI ranges from machine learning to generative AI to virtual assistants to robotics, and this list merely scratches the surface.

When it comes to IP and AI, there are several critical questions for governments to consider: Can AI models be protected by existing legal frameworks within IP? Must copyright owners be human? Does a patent inventor have to be an individual? Do AI models’ training programs infringe on others’ copyrights?

To begin to answer these questions, regulators are drawing from existing IP regimes, including patent and copyright law. Some APAC countries have taken a non-binding approach, relying on existing principles to guide AI regulation. Others are drafting more specific AI regulations. The summary chart below provides a brief overview of current patent and copyright laws within APAC focused on AI and IP. Additional commentary concerning updates to AI laws and regulations is provided below the chart.

Country Patent Copyright
Korea A non-human cannot be the inventor under Korea’s Patent Act. There is a requirement for “a person.” The Copyright Act requires a human creator. Copyright is possible if the creator is a human using generative AI models as software tools and the human input is considered more than simple prompt inputs. For example, in Korea, copyright was granted to a movie produced by generative AI as a “compilation work” in December 29, 2023.
Japan Under Japan’s Patent Act, a natural person must be the inventor. This is the “requirement of shimei 氏名” (i.e. name of a natural person). Japan’s Copyright Act defines a copyright-protected work as “a creation expressing human thoughts and Emotions.” However, in February 29, 2024, the Agency for Cultural Affairs committee’s document on “Approach to AI and Copyright” provided that a joint work made up of both human input and AI generated content can be eligible for copyright protection.
Taiwan Taiwan’s Patent Law does not explicitly preclude a non-human inventor, however, the Patent Examination Guidelines require a natural person to be an inventor. Formalities in Taiwan also require an inventor’s name and nationality. The Copyright Act requires of “human creative expression.”
China The inventor needs to be a person under Patent Law and the Guidelines for Examination in China. Overall, Chinese courts have recognized that when AI-generated works involve human intellectual input, the user of the AI software is the copyright owner.
Hong Kong The Patents Ordinance in Hong Kong requires a human inventor. The Copyright Ordinance in Hong Kong attributes authorship to “the person by whom the arrangements necessary for the creation of the work are undertaken.”
Philippines Patent law in the Philippines requires a natural person to be the inventor. Generally, copyright law in the Philippines requires the author to be a natural person. The copyright in works that are partially AI-generated protects only those parts that are created by natural persons. The Philippines IP Office relies on the declarations of the creator claiming copyright to provide which part of the work is AI-generated and which part is not.
Vietnam AI cannot be an IP right owner in Vietnam. The user of AI is the owner, regardless of the degree of work carried out by AI. In terms of copyright, AI cannot be an IP right owner. Likewise, the user of AI is the owner, regardless of the degree of work carried out by AI.
Thailand Thailan’s Patent law in Thailand requires inventors to be individuals. Copyright law in Thailand requires an author to be an individual.
Malaysia Malaysia’s Patent law requires inventors to be individuals. Copyright law in Malaysia requires an author to be an individual.
Singapore Patent law requires inventors to be a natural person(s). However, the owner can be a natural person or a legal entity. In Singapore, it is implicit in provisions of the Copyright Act that the author must be a natural person.
Indonesia Under Indonesia’s patent law, the inventor may be an individual or legal entity. Under copyright law in Indonesia, the author of a work may be an individual or legal entity.
India India’s patent law requires inventors to be a natural person(s). The copyright law contains a requirement of “originality” – which the courts interpret as “intellectual effort by humans.”
Australia The Full Federal Court in Australia ruled that an inventor must be a natural person. Copyright law in Australia requires the author to be a human.
New Zealand One court in New Zealand has ruled that AI cannot be an inventor under the Patents Act. A court in New Zealand has ruled that AI cannot be the author under the provisions of the Copyright Act. There is updated legislation clarifying that the ownership of computer-generated works is the person who “made the arrangements necessary” for the creation of the work.

AI Regulation and Infringement

KOREA: Court decisions have ruled that web scraping or pulling information from a competitor’s website or database infringes on competitor’s database rights under the Copyright Act and the UCPA. In Koria, parties must obtain permission for use of copyrighted work for training AI emphasized in guidelines. The Copyright Commission published guidelines on copyright and AI in December 2023. The guidelines noted the growing need for legislation on AI generated works. The English version of the guidelines was released in April 2024.

JAPAN: The January 1, 2019 Copyright Act provides very broad rights to use copyrighted works without permission for training AI, as long as the training is for the purpose of technological development. The committee aims to introduce checks to this freedom, and also to provide more protection for Japan-based content creators and copyright holders. The Japan Agency for Cultural Affairs (ACA) released its draft “Approach to AI and Copyright” for public comment on January 23, 2024. Additional changes have been made to the draft after considering 25,000 comments as of February 29, 2025. Also, the Ministry of Internal Affairs and Communications, Ministry of Economy, Trade and compiled the AI Guidelines for Business Ver1.0 in Japan on April 19, 2024.

TAIWAN: Using copyrighted works to train AI models involves “reproduction”, which constitutes an infringement, unless there is consent or a license to use the work. Taiwan’s IPRO released an interpretation to clarify AI issues in June 2023. Under the IPO interpretation circular of June 2023, the Taiwan cabinet approved draft guidelines for the use of generative AI by the executive branch of the Taiwan government in August 2023. The executive branch of the Taiwan government also confirmed that it is in the process of formulating the government’s version of the Draft AI Law, which is expected to be published this year.

CHINA: Interim Measures for the Management of Generative Artificial Intelligence Services, promulgated in July 2023, require that generative AI services “respect intellectual property rights and commercial ethics” and that “intellectual property rights must not be infringed.” The consultation draft on Basic Security Requirements for Generative Artificial Intelligence Service, which was published in October 2023, provides detailed guidance on how to avoid IP infringement. The requirements, for example, provide specific processes concerning model training data that Chinese AI companies must adopt. Moreover, China’s draft Artificial Intelligence Law, proposed on March 16, 2024, outlines the use of copyrighted material for training purposes, and it serves as a complement to China’s current AI regulations.

HONG KONG: A review of copyright law in Hong Kong is underway. There is currently no overarching legislation regulating the use of AI, and the existing guidelines and principles mainly provide guidance on the use of personal data.

VIETNAM: AI cannot have responsibility for infringement, and there are no provisions under existing laws in Vietnam regarding the extent of responsibility of AI users for infringing acts. The Law on Protection of Consumers’ Rights will take effect on July 1, 2024. This law requires operators of large digital platforms to periodically evaluate the use of AI and fully or partially automated solutions.

THAILAND: Infringement in Thailand requires intent or implied intent, for example, from the prompts made to the AI. Thai law also provides for liability arising out of the helping or encouraging of infringement by another. Importantly, the AI user may also be exposed to liability in that way.

MALAYSIA: An informal comment from February 2024 by the Chairman of the Malaysia IP Office provides that there may be infringement through the training and/or use of AI programs.

SINGAPORE: Singapore has a hybrid regime. The regime provides a general fair use exception, which is likely guided by US jurisprudence, per the Singapore Court of Appeal. The regime also provides exceptions for specific types of permitted uses, for example, the computational data analysis exception. A Landscape Report on Issues at the Intersection of AI and IP issued by IPOS on February 28, 2024 provided a Model AI Governance Framework for Generative AI, which was published May 30, 2024.

INDONESIA: A “circular,” a government issued document similar to a white paper, implies that infringement is possible in Indonesia. The nonbinding Communications and Information Ministry Circular No. 9/2023 on AI was signed in December 2023.

INDIA: Under the Copyright Act of 1957, a Generative AI user has an obligation to obtain permission to use the copyright owner’s works for commercial purposes. In February 2024, the Ministry of Commerce and Industry’s Statement provided that India’s existing IPR regime is “well-equipped to protect AI-generated works” and therefore, it does not require a separate category of rights. MeitY issued a revised advisory on March 15, 2024 providing that platforms and intermediaries should ensure that the use of AI models, large language models, or generative AI software or algorithms by end users does not facilitate any unlawful content stipulated under Rule 3(1)(b) of the IT Rules, in addition to any other laws.

AUSTRALIA: Any action seeking compensation for infringement of a copyright work by an AI system would need to rely on the Copyright Act of 1968. It is an infringement of copyright to reproduce or communicate works digitally without the copyright owner’s permission. Australia does not have a general “fair use” defense to copyright infringement.

NEW ZEALAND: While infringement by AI users has not yet considered by New Zealand courts, New Zealand has more restricted “fair dealing” exceptions. Copyright review is underway in New Zealand.

Arguing Internet Availability to Establish Copyright Infringement Is Bananas

In an unpublished opinion, the US Court of Appeals for the Eleventh Circuit affirmed a district court’s decision finding that a pro se Californian artist failed to establish that an Italian artist had reasonable opportunity to access the copyrighted work simply because it was available to view on the internet. Morford v. Cattelan, Case No. 23-12263 (11th Cir. Aug. 16, 2024) (Jordan, Pryor, Branch, JJ.) (per curiam).A plaintiff alleging copyright infringement may show factual copying by either direct or indirect evidence showing “that the defendant had access to the copyrighted work and that there are probative similarities between the allegedly infringing work and the copyrighted work.” To do so, however, the copyright owner must establish a nexus between the work and the defendant’s alleged infringement. Mere access to a work disseminated in places or settings where the defendant may have come across it is not sufficient.

Joe Morford’s Banana and Orange and Maurizio Cattelan’s Comedian both “involve the application of duct tape to a banana against a flat surface” (see images below from the court decision’s appendix). Cattelan’s Comedian went viral and sold for more than $100,000 at Miami’s Art Basel. Morford claimed that Comedian was a copy. The district court found that Morford failed to show that Cattelan had reasonable opportunity to access Banana and Orange and thus could not establish a copyright claim. Morford appealed.

Orange and Banana, Comedian

On appeal, Morford argued that because he could show striking similarity between Banana and Orange and Comedian, he was not required to proffer evidence of access to show copyright infringement. In the alternative, he argued that he could show substantial similarity and that Cattelan had reasonable opportunity to access Banana and Orange as it was widely disseminated and readily discoverable online.

The Eleventh Circuit explained that in circuits adopting a widespread dissemination standard, that standard requires showing that the work enjoyed “considerable success or publicity.” Morford showed that Banana and Orange was available on his public Facebook page for almost 10 years and featured on his YouTube channel and in a blog post, with views in more than 25 countries. But Banana and Orange’s availability on the internet, without more, was “too speculative to find a nexus” between Cattelan and Morford to satisfy the factual copying prong of a copyright infringement claim, according to the Court.

The Eleventh Circuit also found that Morford failed to meet the high burden of demonstrating that the original work and accused infringement were so strikingly similar as to establish copying. Such similarity exists if the similarity in appearance between the two works “is so great that [it] precludes the possibility of coincidence, independent creation or common source,” but identical expression does not necessarily constitute infringement. In this analysis, a court addresses the “uniqueness or complexity of the protected work as it bears on the likelihood of copying.” Morford argued that he established striking similarity based on the “same two incongruous items being chosen, grouped, and presented in the same manner within both works.” Although the two incongruous items in both works were similar (i.e., a banana and duct tape), the Court decided that there were sufficient differences between Banana and Orange and Comedian to preclude a finding of striking similarity. Banana and Orange had both a banana and an orange held by duct tape, while Comedian only contained a banana.

AI-Generated Content and Trademarks

The rapid evolution of artificial intelligence has undeniably transformed the digital landscape, with AI-generated content becoming increasingly common. This shift has profound implications for brand owners introducing both challenges and opportunities.

One of the most pressing concerns is trademark infringement. In a recent example, the Walt Disney Company, a company fiercely protective of its intellectual property, raised concerns about AI-generated content potentially infringing on its trademarks.  Social media users were having fun using Microsoft’s Bing AI imaging tool, powered by DALL-E 3 technology, to create images of pets in a “Pixar” style.  However, Disney’s concern wasn’t the artwork itself, but the possibility of the AI inadvertently generating the iconic Disney-Pixar logo within the images, constituting a trademark infringement. This incident highlights the potential for AI-generated content to unintentionally infringe upon established trademarks, requiring brand owners to stay vigilant in protecting their intellectual property in the digital age.

Dilution of trademarks is another critical issue. A recent lawsuit filed by Getty Images against Stability AI sheds light on this concern. Getty Images, a leading provider of stock photos, accused Stability AI of using millions of its copyrighted images to train its AI image generation software. This alleged use, according to Getty Images, involved Stability AI’s incorporation of Getty Images’ marks into low-quality, unappealing, or offensive images which dilutes those marks in further violation of federal and state trademark laws. The lawsuit highlights the potential for AI, through the sheer volume of content it generates, to blur the lines between inspiration and infringement, weakening the association between a trademark and its source.

In addition, the ownership of copyrights in AI-generated marketing can cause problems. While AI tools can create impressive content, questions about who owns the intellectual property rights persist.  Recent disputes over AI-generated artwork and music have highlighted the challenges of determining ownership and copyright in this new digital frontier.

However, AI also presents opportunities for trademark owners. For example, AI can be employed to monitor online platforms for trademark infringements, providing an early warning system. Luxury brands have used AI to authenticate products and combat counterfeiting. For instance, Entrupy has developed a mobile device-based authentication system that uses AI and microscopy to analyze materials and detect subtle irregularities indicative of counterfeit products. Brands can integrate Entrupy’s technology into their retail stores or customer-facing apps.

Additionally, AI can be a powerful tool for brand building. By analyzing consumer data and preferences, AI can help create highly targeted marketing campaigns. For example, cosmetic brands have successfully leveraged AI to personalize product recommendations, enhancing customer engagement and loyalty.

The intersection of AI and trademarks is a dynamic and evolving landscape. As technology continues to advance, so too will the challenges and opportunities for trademark owners. Proactive measures, such as robust trademark portfolios, AI-powered monitoring tools, and clear internal guidelines, are essential for safeguarding brand integrity in this new era.

Relying on Noncompete Clauses May Not Be the Best Defense of Proprietary Data When Employees Depart

Much of the value of many companies often is wrapped up with and measured by their intellectual property (IP) portfolios. Some forms of IP, such as patents, are known by the public. Others derive their value from being hidden from the public. Many companies, for example, have gigabytes of data or “know-how” that may be worth millions, but only to the extent that they remain secret. This article discusses some ways to keep business information confidential when an employee who has had access to that information leaves the company.

Many companies traditionally turned to employment agreements, specifically noncompete clauses, to protect proprietary competitive information. The legality of noncompetes is in question following the Federal Trade Commission’s (FTC’s) ban on them, which is being challenged in court by the U.S. Chamber of Commerce, causing confusion and concerns about protecting information via noncompete agreements. As covered in Wilson Elser’s prior articles* on this subject, the timeline of the FTC rule in question was as follows:

  • The FTC promulgated new rules to take effect in September 2024 banning all noncompete agreements.
  • The U.S. Supreme Court overturned the 40-year-old method of reviewing agency rules (Chevron Deference), throwing all agency rules, including the FTC’s rule on noncompetes, into question.
  • The District Court for the Northern District of Texas preliminarily enjoined the FTC from enforcing its new rule banning noncompetes.

After this flurry of activity, noncompetes are, for now, not banned. But do they offer an effective solution for businesses seeking to protect their proprietary information?

Noncompete Clauses Are Not Always Effective
Vortexa, Inc. v. Cacioppo, a June 2024 case from the District Court for the Southern District of New York, illustrates the limitations of noncompete clauses in employment agreements. That case presents the familiar fact pattern of an employee leaving and going to work for a competitor. With some evidence of the employee’s access to proprietary competitive information in hand (but no evidence of actual misappropriation), the former employer sought a preliminary injunction to prevent the employee from working for the competitor for one year, the term stated in the noncompete clause in the employee’s contract with the former employer. The contract also included common non-disclosure and confidentiality clauses.

Absent evidence of actual misappropriation, the plaintiff employer relied on the “Inevitable Disclosure” doctrine, which assumes that a departing employee will inevitably disclose confidential information when they go work for a competitor. The court refused to apply this doctrine, explaining that inevitable disclosure may substitute for actual evidence of misappropriation only when the information is a trade secret. Here, none of the information about which the former employer was concerned was a trade secret.

The proprietary information that the former employee had was pricing data, marketing strategies and “intricacies of the business.” These types of information do not, in and of themselves, constitute trade secrets. In addition, the information was not afforded trade secret treatment because (1) some of it was ascertainable by the competitor without reference to the first employer’s information; (2) the companies sell different products; (3) some of the information was developed without the expenditure of a good deal of money and effort; (4) some of the information was provided to clients without a non-disclosure agreement; (5) some of the information was shared on company-wide collaboration channels; and (6) “google drive log records show that [the former employee] opened and viewed these documents, which underlines the lack of security protecting this purportedly confidential information.”

Most of these reasons for the information not being accorded trade secret status cannot be changed by any action of the employer. For example, if information can be generated by means independent of the first employer, that information cannot be protected by trade secret law and nothing the first employer can do will change that after the fact. However, any business seeking to protect its valuable competitive information can change the way that it secures, protects and manages access to its competitive information, and this may be enough to ensure that its information is protected by trade secret law.

What Businesses Should Do to Protect Their Proprietary Competitive Information
Generally, proprietary competitive information can be protected as a trade secret by operation of law or via contract. In many cases, the “boots and suspenders” approach is best – the information should be protected both by contract and by meeting the requirements for protection under trade secret law. As described, a contract alone is sometimes ineffective, so information that derives its value from not being generally known to the public should also be treated in such a manner that the courts would see it as being a trade secret.

Specifically, for something to qualify for trade secret protection under federal and state statues and common law, it must be securely kept and carefully protected from disclosure. Some easy ways to protect information are to (1) restrict access to folders on a company’s internal computer systems, (2) physically lock rooms that contain hard copies and (3) have computers lock automatically when not accessed for set time periods. Protecting information via noncompete, confidentiality and non-disclosure contractual obligations is another way to ensure that information remains secret, such that it is protected under trade secret law. Internal policies on how information may be shared with third parties, such as clients, also are helpful evidence of trade secret treatment. In addition, the business may consider maintaining records on the time, effort and monetary expenditures required to develop proprietary information, which should allow the business to demonstrate that making such information freely available to a competitor is fundamentally unfair.

In some cases, information protected as a trade secret may be the most valuable IP that a company owns. But the value can easily be lost if the company does not properly secure the information. Different scenarios call for different methods of security, and a good rule of thumb to protect information from disclosure by a departing employee is to protect this information both by contract and as a trade secret.

The first step for any business is to think through their overall data protection strategy and consult with experienced intellectual property counsel to put appropriate protections in place.

Fourth Circuit Reverses $1 Billion Award for Vicarious Liability Claim for More than 10,000 Works

On January 12, 2021, the U.S. District Court for the Eastern District of Virginia awarded a group of music recording companies (the plaintiffs) a $1 billion verdict against Cox Communications (Cox). The Virginia court’s ruling found that Cox, an internet service provider (ISP), was contributorily and vicariously liable for copyright infringement committed by certain subscribers on its networks. The plaintiffs alleged that the ISP allowed the unauthorized downloading and distribution of more than 10,000 copyrighted works by Cox subscribers who had already received three or more notices of infringement. The district court in Virginia established that the “takedown” notices sent by the plaintiffs provided Cox with the requisite knowledge of its subscribers’ repeated infringement to substantiate their claim that Cox was contributorily liable, suggesting that Cox had sufficient specific knowledge of infringement to have done something about it.

The plaintiffs’ notice to Cox identified the IP address of the subscriber, as well as the time of infringement and the identification of the infringed work, which the plaintiffs argued was sufficiently specific knowledge for Cox to be able to identify the subscriber and to exercise its policy by suspending or terminating the infringing subscriber. This case proceeded to trial on two theories of secondary liability – vicarious and contributory copyright infringement. The plaintiffs argued that Cox failed to act on these known repeat infringers, and the jury found Cox liable for willful contributory infringement and vicarious infringement, ordering Cox to pay more than $99,000 for each of the infringed-upon works. Cox appealed the jury verdict.

On appeal, before the U.S. Court of Appeals for the Fourth Circuit, Cox raised several questions of law concerning the secondary liability for copyright infringement, as well as what constitutes a derivative work in the Internet Age.

Vicarious Infringement
The Fourth Circuit’s analysis first considered whether the district court erred in denying plaintiffs’ vicarious infringement claim. “A defendant may be held vicariously liable for a third party’s copyright infringement [if the defendant] (1) profits directly from the infringement and (2) has a right and ability to supervise the direct infringer.” See Metro-Goldwyn-Mayer Studios, Inc. v. Grokster, Ltd., 545 U.S. 913, 930 n.9 (2005) (internal citations omitted). The Fourth Circuit found that the plaintiffs failed to establish the first element as a matter of law and thus found that the plaintiffs failed to establish that Cox was vicariously liable.

In reaching this decision, the Fourth Circuit turned to the landmark decision in Shapiro, Bernstein & Co., 316 F.2d 304 (2d Cir. 1963), a case on vicarious liability for infringing copyrighted music recordings. In Shapiro, a department store was sued for the selling of “bootleg” records by a concessionaire operating in its stores. The store had the right to supervise the concessionaire and employees, demonstrating its control over the infringement. There, the store received a certain percentage of every record sale, “whether ‘bootleg’ or legitimate,” giving it “a more definite financial interest” in the infringing sales.” Thus, the Shapiro court found that the financial gains were clearly spelled out from the bootleg sales and acts of infringement in Shapiro.

Next, the Fourth Circuit recognized that courts have found that a defendant may possess a financial interest in a third party’s infringement of copyrighted music, even absent a strict correlation between each act of infringement and an added penny of profits. See Fonovisa, Inc. v. Cherry Auction, Inc., 76 F.3d 259 (9th Cir. 1996). In Fonovisa, the operator of a swap meet allowed vendors to sell infringing goods, and the operator collected “admission fees, concession stand sales, and parking fees” but no sales commission “from customers who want[ed] to buy the counterfeit recordings at bargain-basement prices.” The Fonovisa court found that the plaintiffs adequately showed a financial benefit from the swap meet owner and the sales of pirated recordings at the swap meet, which was a draw for customers. Thus, the infringing sales “enhance[d] the attractiveness of the venue of the potential customers, finding the swap meet operator had a financial interest in the infringement sufficient to state a claim for vicarious liability.”

The Fourth Circuit established that Shapiro and Fonovisa provided the steppingstones of the principles of copyright infringement to the internet and cyberspace and that Congress agreed that “receiving a one-time setup fee and flat periodic payment for service” from infringing and non-infringing users alike ordinarily “would not constitute a financial benefit directly attributable to the infringing activity.” Ellison v. Robertson, 357 F. 3d 1072, 1079 (9th Cir. 2004) (internal citations omitted). The Court also reviewed other court precedents, including A&M Records v. Napster, Inc., 239 F.3d 1004 (9th Cir. 2001), to show that increased pirated music drew in users as a direct financial interest for vicarious liability., but also notes that courts have found no evidence of a direct financial benefit between subscribers of American Online (AOL) and the availability of infringing content.’’ Ellison, 357 F.3d at 1079.

Against this backdrop, the Fourth Circuit held that to prove Cox was vicariously liable, the plaintiffs had to demonstrate that Cox profited from its subscribers’ infringing download and distribution of the plaintiffs’ copyrighted songs, which – given the evidence at trial – it did not. While the district court found it was enough that Cox repeatedly declined to cancel an ISP subscriber’s monthly subscription fee, the Fourth Circuit found this evidence to be insufficient. Instead, the Fourth Circuit found that the continued monthly payment fees for internet service, even by repeat infringers, was not a financial benefit flowing directly from the copyright infringement. Cox established that subscribers paid a flat fee even if all of its subscribers stopped infringing. Recognizing that an internet provider would necessarily lose money if it canceled subscriptions only demonstrates that service providers have a direct financial interest in providing subscribers with access to the internet only. Thus, the Fourth Circuit held that vicarious liability demands proof that the defendant profits directly from the acts of infringement for which it is being held accountable.

To rebut this, the plaintiffs claimed that the jury could infer that subscribers paid monthly membership fees based on the high volume of infringing content. The Fourth Circuit rejected this argument and found that the evidence was insufficient to prove that customers were drawn to Cox’s internet service or that they continued the service because they were specifically drawn to the opportunity to infringe the plaintiffs’ copyrights. The plaintiffs further asserted that subscribers were willing to pay more for the opportunity to infringe based on Cox’s tiered structure for internet access – but the plaintiffs fell short in proving this claim because no reasonable inference could be drawn that Cox subscribers paid more for faster internet to infringe on the copyrighted works. Ultimately, the Court found that the plaintiffs could not establish a causal connection between subscribers’ copyright infringement and Cox’s revenue for monthly subscriptions. Thus, the Fourth Circuit held that Cox was not liable for its subscribers’ copyright infringement and reversed the district court’s ruling on this theory. The court vacated the $1 billion damages award and remanded the case for a new trial on damages, holding that the jury’s finding of vicarious liability could have influenced its assessment of statutory damages.

Contributory Infringement
The Fourth Circuit then examined the remaining issue of contributory infringement. Under this theory, “one who, with knowledge of the infringing activity, induces, causes or materially contributes to the infringing conduct of another is liable for the infringement, too.” Cox argued that the district court erred by taking away the factual determination from the jury that notices of past infringement established Cox’s knowledge that subscribers were substantially certain to infringe in the future. Cox had contracted with a third party to provide copyright violation notices to users and asserted that it used these notices as their safe harbor under the Digital Millennium Copyright Act to alert violators and to terminate access to users who were repeat infringers. Despite this, the Fourth Circuit ultimately agreed with the jury’s finding that Cox materially contributed to copyright infringement occurring on its network and that its conduct was culpable.

Therefore, a three-judge panel found that Cox was liable for willful copyright infringement but reversed the vicarious liability verdict and remanded a new trial on damages. The Fourth Circuit held that because Cox did not profit from its subscribers’ acts of infringement, a legal prerequisite for vicarious liability, Cox was not liable for damages under the vicarious liability theory.

The Impact
The Fourth Circuit’s decision recognizes a new dawn breaking in copyright law, one that requires a causal connection between profit and/or financial gain and a defendant’s acts of infringement to prove vicarious liability in a copyright infringement claim under the Copyright Act. The plaintiffs attempted to bridge the financial gap between acknowledging access to infringing content through a monthly internet subscription and high-volume infringing acts. However, the Fourth Circuit found that this leap in logic was a step too far and reversed the award for vicarious liability for lack of evidence to find this missing connection between Cox subscribers and infringing plaintiffs’ content.

While this may be one route the courts may consider to reduce music piracy damages, it remains to be seen whether other courts will take this approach to determining that profit is the key element supporting other vicarious liability claims in cyberspace.

Huawei U.S. Patent Grants Drop 24% in 2023; BOE Drops Out of the Top 10

Accordingly to analysis by Harrity Patent AnalyticsHuawei ranked 10th for US patent grants in 2023, down 3 spots from 2022 with a 24% drop in patent grants. BOE Technology Group Co., Ltd. dropped even further to 19th down from 8th in 2022. BOE had a 38% drop in granted US patents. Tencent was 54th this year with an 11% drop in patent grants. Oppo was down 32%. Baidu was up only 1%. Xiaomi was down 7% and didn’t make the top 100 in 2023. Alibaba was down 49%. In contrast, US companies advanced with Qualcomm patent grants up 46%, Alphabet (Google) up 23% and Apple up 11%.

The drop in Chinese patent grants reverse the trend of growing US patent grants for Chinese companies. Huawei’s patents grants in 2022 were up 3%, BOE up 27%, Baidu up 43% and Xiaomi up 33%.

Possible reasons for the drop in grants to Chinese entities may include a poorly performing Chinese economy, the reduction and elimination of government subsidies for foreign patent grants, the impact of COVID-19 on patent application filings the past few years; and geopolitical tensions.

Top 10 Chinese Grantees of U.S. Patents in 2023

Rank Company 2023 Patents % Change from 2022
10 HUAWEI TECHNOLOGIES CO., LTD. 2290 -24%
19 BOE TECHNOLOGY GROUP CO., LTD 1695 -38%
54 TENCENT HOLDINGS LTD 702 -11%
67 BAIDU, INC. 626 +1%
77 LENOVO GROUP LIMITED 530 -16%
84 OPPO MOBILE TELECOMMUNICATIONS CORPORATION 516 -32%
101 XIAOMI INC. 461 -7%
115 TSINGHUA HOLDINGS 372 -3%
121 ZTE CORPORATION 351 -14%
122 BYTEDANCE LTD. 350 +119%

Source: https://harrityllp.com/patent300/

To AI or Not to AI: U.S. Copyright Office Clarifies Options

The U.S. Copyright Office has weighed in with formal guidance on the copyrightability of works whose generation included the use of artificial intelligence (AI) tools. The good news for technology-oriented human creative types: using AI doesn’t automatically disqualify your work from copyright protection. The bad news for independent-minded AI’s: you still don’t qualify for copyright protection in the United States.

On March 16, 2023, the Copyright Office issued a statement of policy (“Policy”) to clarify its practices for examining and registering works that contain material generated by the use of AI and how copyright law’s human authorship requirements will be applied when AI was used. This Policy is not itself legally binding or a guarantee of a particular outcome, but many copyright applicants may breathe a sigh of relief that the Copyright Office has formally embraced AI-assisted human creativity.

The Policy is just the latest step in an ongoing debate over the copyrightability of machine-assisted products of human creativity. Nearly 150 years ago, the Supreme Court ruled at photographs are copyrightable. See Burrow-Giles Lithographic Company v. Sarony, 111 U.S. 53 (1884). The case involved a photographer’s claim against a lithographer for 85,000 unauthorized copies of a photograph of Oscar Wilde. The photo, Sarony’s “Oscar Wilde No. 18,” is shown below:

Sarony’s “Oscar Wilde No. 18"

The argument against copyright protection was that a photograph is “a reproduction, on paper, of the exact features of some natural object or of some person” and is therefore not a product of human creativity. Id. at 56. The Supreme Court disagreed, ruling that there was sufficient human creativity involved in making the photo, including posing the subject, evoking the desired expression, arranging the clothing and setting, and managing the lighting.

In the mid-1960’s, the Copyright Office rejected a musical composition, Push Button Bertha, that was created by a computer, reasoning that it lacked the “traditional elements of authorship” as they were not created by a human.

In 2018, the U.S. Court of Appeals for the Ninth Circuit ruled that Naruto, a crested macaque (represented by a group of friendly humans), lacked standing under the Copyright Act to hold a copyright in the “monkey selfie” case. See Naruto v. Slater, 888 F.3d 418 (9th Cir. 2018). The “monkey selfie” is below:

Monkey Selfie

In February 2022, the Copyright Office rejected a registration (filed by interested humans) for a visual image titled “A Recent Entrance to Paradise,” generated by DABUS, the AI whose claimed fractal-based inventions are the subject of patent applications around the world. DABUS’ image is below:

“A Recent Entrance to Paradise”

Litigation over this rejected application remains pending.

And last month, the Copyright Office ruled that a graphic novel consisting of human-authored text and images generated using the AI tool Midjourney could, as a whole, be copyrighted, but that the images, standing alone, could not. See U.S. Copyright Office, Cancellation Decision re: Zarya of the Dawn (VAu001480196) at 2 (Feb. 21, 2023).

The Copyright Office’s issuing the Policy was necessitated by the rapid and remarkable improvements in generative AI tools over even the past several months. In December 2022, generative AI tool Dall-E generated the following images in response to nothing more than the prompt, “portrait of a musician with a hat in the style of Rembrandt”:

Four portraits generated by AI tool Dall-E from the prompt, "portrait of a musician with a hat in the style of Rembrandt."

If these were human-generated paintings, or even photographs, there is no doubt that they would be copyrightable. But given that all four images were generated in mere seconds, with a single, general prompt from a human user, do they meet the Copyright Office’s criteria for copyrightability? The answer, now, is a clear “no” under the Policy.

However, the Policy opens the door to registering AI-assisted human creativity. The toggle points will be:

“…whether the ‘work’ is basically one of human authorship, with the computer [or other device] merely being an assisting instrument, or whether the traditional elements of authorship in the work (literary, artistic, or musical expression or elements of selection, arrangement, etc.) were actually conceived and executed not by man but by a machine.” 

In the case of works containing AI-generated material, the Office will consider whether the AI contributions are the result of “mechanical reproduction” or instead of an author’s “own original mental conception, to which [the author] gave visible form.” 

The answer will depend on the circumstances, particularly how the AI tool operates and how it was used to create the final work. This will necessarily be a case-by-case inquiry.” 

See Policy (citations omitted).

Machine-produced authorship alone will continue not to be registerable in the United States, but human selection and arrangement of AI-produced content could lead to a different result according to the Policy. The Policy provides select examples to help guide registrants, who are encouraged to study them carefully. The Policy, combined with near future determinations by the Copyright Office, will be critical to watch in terms of increasing likelihood a registration application will be granted as the Copyright Office continues to assess the impacts of new technology on the creative process. AI tools should not all be viewed as the “same” or fungible. The type of AI and how it is used will be specifically considered by the Copyright Office.

In the short term, the Policy provides some practical guidance to applicants on how to describe the role of AI in a new copyright application, as well as how to amend a prior application in that regard if needed. While some may view the Policy as “new” ground for the Copyright Office, it is consistent with the Copyright Office’s long-standing efforts to protect the fruits of human creativity even if the backdrop (AI technologies) may be “new.”

As a closing note, it bears observing that copyright law in the United Kingdom does permit limited copyright protection for computer-generated works – and has done so since 1988. Even under the U.K. law, substantial questions remain; the author of a computer-generated work is considered to be “the person by whom the arrangements necessary for the creation of the work are undertaken.” See Copyright, Designs and Patents Act (1988) §§ 9(3), 12(7) and 178. In the case of images generated by a consumer’s interaction with a generative AI tool, would that be the consumer or the generative AI provider?

Copyright © 2023 Womble Bond Dickinson (US) LLP All Rights Reserved.

Breaking News – Hermès Makes History With First NFT Trademark Trial Victory

A New York City jury just returned a verdict in favor of Hermès in a historic dispute between the luxury fashion house and digital artist Mason Rothschild over Hermès’ alleged trademark rights relating to Hermès’ famous Birkin handbag. The jury awarded Hermès $133,000 in total damages for trademark infringement, dilution, and cybersquatting.

The jury finding that the First Amendment did not shield Rothschild from liability in connection with his MetaBirkins NFTs project is significant, particularly as this matter involved the first trial by jury to consider the interplay of free speech and trademark protection in the context of NFTs. This decision, which may be appealed, provides guidance for artists, brands, and others seeking ingress into metaverse, including to what extent “real world” intellectual property rights apply to and may be enforced in virtual worlds.

Haute-ly Contested NFTs

Throughout the dispute over this past year, the parties have contested each other’s characterization of the MetaBirkins NFTs. To Hermès, the MetaBirkins NFTs are merely the instruments of a “digital speculator” looking to exploit one of its most exclusive assets via NFTs. In contrast, Rothschild argues that the MetaBirkins NFTs project, a series of 100 NFT images that depict a range of reimagined Hermès Birkin bags featuring a variety of colorful fur, is digital art and a commentary on the famed BIRKIN bag, consumerism, and animal cruelty within the fashion industry. As a result, he argues that the MetaBirkins NFTs are artistic works that should be shielded from liability under the free speech principles of the First Amendment of the Constitution. The nine-member jury disagreed, finding that the MetaBirkins NFTs were more like commodities that are subject to trademark and other laws, rather than artwork. A factor that may have influenced the jury’s decision was evidence suggesting that Rothschild may have seen the MetaBirkins NFTs as a “cash cow.” This may have cast doubt on the authenticity of his characterization of the MetaBirkins NFTs as an art project.

The Test is Yet to Come

Although the jury found the MetaBirkins NFTs to be infringing, the final disposition of this dispute remains pending with the possibility of appeal. Given the importance of the issues at stake, the outcome of this case is bound to be subject to debate regardless of any appeal.

Moreover, while no NFT-specific legal test appears to have emerged from this case and the legal landscape for IP in the Metaverse (and beyond) continues to lack clear guidance, this case has nonetheless provided insight on how courts (and juries) may view the interplay of IP and NFTs. The ultimate outcome of this landmark case is likely to form the basis of the emerging law involving IP rights and NFTs.

© 2023 ArentFox Schiff LLP
For more Intellectual Property Legal News, click here to visit the National Law Review

With the US Copyright Office (USCO) continuing their stance that protection only extends to human authorship, what will this mean for artificial intelligence (AI)-generated works — and artists — in the future?

Almost overnight, the limited field of Machine Learning and AI has become nearly as accessible to use as a search engine. Apps like Midjourney, Open AI, ChatGPT, and DALL-E 2, allow users to input a prompt into these systems and a bot will generate virtually whatever the user asks for. Microsoft recently announced its decision to make a multibillion-dollar investment in OpenAI, betting on the hottest technology in the industry to transform internet as we know it.[1]

However, with accessibility of this technology growing, questions of authorship and copyright ownership are rising as well. There remain multiple open questions, such as: who is the author of the work — the user, the bot, or the software that produces it? And where is this new generative technology pulling information from?

AI and Contested Copyrights

As groundbreaking as these products are, there has been ample backlash regarding copyright infringement and artistic expression. The stock image company, Getty Images, is suing Stability AI, an artificial intelligence art tool behind Stable Diffusion. Getty Images alleges that Stability AI did not seek out a license from Getty Images to train its system. Although the founder of Stability AI argues that art makes up 0.1% of the dataset and is only created when called by the user’s prompt. In contrast, Shutterstock, one of Getty Images largest competitors, has taken an alternative approach and instead partnered with Open AI with plans to compensate artists for their contributions.

Artists and image suppliers are not the only ones unhappy about the popularity of machine learning.  Creators of open-source code have targeted Microsoft and its subsidiary GitHub, along with OpenAI,  in a proposed class-action lawsuit. The lawsuit alleges that the creation of AI-powered coding assistant GitHub Copilot is relying on software piracy on an enormous scale. Further, the complaint claims that GitHub relies on copyrighted code with no attribution and no licenses. This could be the first class-action lawsuit challenging the training and output of AI systems. Whether artists, image companies, and open-source coders choose to embrace or fight the wave of machine learning,  the question of authorship and ownership is still up for debate.

The USCO made clear last year that the copyright act only applies to human authorship; however they have recently signaled that in 2023 the office will focus on the legal grey areas surrounding the copyrightability of works generated in conjunction with AI. The USCO denied multiple applications to protect AI authored works previously, stating that the “human authorship” element was lacking. In pointing to previous decisions, such as the 2018 decision that a monkey taking a selfie could not sue for copyright infringement, the USCO reiterated that “non-human expression is ineligible for copyright protection.” While the agency is standing by its conclusion that works cannot be registered if it is exclusively created by an AI, the office is considering the issue of copyright registration for works co-created by humans and AI.

Patent Complexities  

The US Patent and Trademark Office (USPTO) will have to rethink fundamental patent policies with the rise of sophisticated AI systems as well. As the USPTO has yet to speak on the issue, experts are speculating alternative routes that the office could choose to take: declaring AI inventions unpatentable, which could lead to disputes and hinder the incentive to promote innovation, or concluding that the use of AI should not render otherwise patentable inventions unpatentable, but would lead to complex questions of inventorship. The latter route would require the USPTO to rethink their existing framework of determining inventorship by who conceived the invention.

Takeaway

The degree of human involvement will likely determine whether an AI work can be protected by copyright, and potentially patents. Before incorporating this type of machine learning into your business practices, companies should carefully consider the extent of human input in the AI creation and whether the final work product will be protectable. For example:

  • An apparel company that uses generative AI to create a design for new fabric may not have a protectable copyright in the resulting fabric design.

  • An advertising agency that uses generative AI to develop advertising slogans and a pitch deck for a client may not be able to protect the client from freely utilizing the AI-created work product.

  • A game studio that uses generative AI to create scenes in a video game may not be able to prevent its unlicensed distribution.

  • A logo created for a business endeavor may not be protected unless there are substantial human alterations and input.

  • Code that is edited or created by AI may be able to be freely copied and replicated.

Although the philosophical debate is only beginning regarding what “makes” an artist, 2023 may be a uniquely litigious year defining the extent in which AI artwork is protectable under existing intellectual property laws.


FOOTNOTES

[1] https://www.cnn.com/2023/01/23/tech/microsoft-invests-chatgpt-openai/index.htmlhttps://www.nytimes.com/2023/01/12/technology/microsoft-openai-chatgpt.html

IP Rights in Virtual Fashion: Lessons Learned in 2022 and Unanswered Questions

There was a lot of talk and much hype about the “metaverse” in 2022. While some were skeptical and stayed on the sidelines to watch, many companies began offering virtual counterparts to their real-world products for use by avatars in the metaverse, including virtual clothing and accessories. For example, Tommy Hilfiger live-streamed a virtual fashion show on Roblox as part of the New York Fashion Week, and Decentraland hosted a Metaverse Fashion Week. Many companies also introduced NFTs into fashion product lines, such as Alo’s NFT offering.

The emergence of virtual goods has generated novel questions about how to protect and enforce IP rights in virtual fashion, and how those strategies might differ from IRL (meaning “in real life”) fashion. Although many questions remain unanswered, this article sets out important considerations for how companies might use various IP laws to protect virtual fashion goods in the United States.

I. DISTINCTIONS BETWEEN VIRTUAL FASHION AND IRL FASHION

Before diving into the IP discussion, it’s worth highlighting some distinctions between virtual fashion and IRL fashion outside the legal context, beyond the obvious fact that virtual fashion is worn by avatars. IRL clothing and accessories are worn primarily for protection against the elements, to conform to societal standards, to conform with a specific event’s dress requirements, to communicate via express messages on clothing or accessories, or to express oneself through the style or design of the clothing.

Virtual fashion can also serve each of those purposes for an avatar, and in some cases the person behind the avatar. But, because it is comprised of software code, the possibilities for virtual fashion utility are endless. For example, a particular piece of virtual clothing can also grant access to certain virtual spaces or events or give the avatar special powers within virtual worlds. If tied to an NFT (non-fungible token), virtual clothing can also provide benefits on and off virtual platforms, including exclusive access to sales promotions and IRL events.

Unlike IRL clothing, however, virtual fashion items currently face compatibility limitations, as the ability to use any virtual fashion item across all virtual platforms is unlikely.

To muddy the waters, as virtual and augmented reality technologies are becoming more popular, they can blur the lines between IRL and virtual fashion. For example, an IRL sweatshirt, when viewed through an appropriate lens, could feature virtual components.

II. IP PROTECTION FOR VIRTUAL FASHION

Because there are no IP laws specific to virtual fashion items, we must seek protection from laws that have traditionally applied to real-life clothing, namely, trademark, trade dress, copyright, and design patent. But the application of these laws can sometimes differ in the virtual context. Each is addressed below.

A. TRADEMARK

Trademark law protects source identifiers such as words, names, logos, and slogans. Obtaining trademark rights specifically in virtual goods, whether acquired through use in commerce or federal registration, is generally straightforward and similar to marks covering IRL fashion. This is evidenced by many marks that were registered in 2022 and specifically cover virtual goods.

That said, even if a company does not have trademark coverage specifically for its virtual goods, the owner of a trademark covering IRL fashion items should have strong arguments that such trademark rights extend to their virtual counterparts. To that point, the U.S. Patent & Trademark Office (USPTO) has refused registration of marks covering virtual goods and services based on prior registrations for the identical marks covering the corresponding IRL goods and services. See, e.g., the refusals of Application No. 97112038 for the mark GUCCI and Application No. 97112054 for the mark PRADA, each of which were filed by parties unrelated to the famous brands.

However, for purposes of enforcement outside of the USPTO context, if a defendant’s goods are virtual, it would have a stronger argument that such goods are not commercial products, but rather expressive works protected by the First Amendment. If a court accepts such an argument, it must then weigh the plaintiff’s trademark rights against the defendant’s First Amendment right of free expression, meaning it would be more challenging for a brand owner to enforce its trademark rights.

In this regard, please see our earlier alert regarding the Hermès v. Rothschild case, in which the court deemed NFTs tied to images of bags called “MetaBirkins” subject to First Amendment protection. [1] In denying Rothschild’s motion to dismiss, the court acknowledged in a footnote that virtually wearable bags (i.e., as opposed to virtual fashion that is displayable but not wearable) might not be afforded First Amendment protection. But we suspect defendants will argue even virtually wearable items should be afforded First Amendment protection, especially given that video games have received such protection. [2]

On balance, companies should consider seeking federal trademark registration specifically for virtual goods and services, for a few reasons:

More direct coverage could help a company in an enforcement action against infringing virtual goods, even if the defendant successfully argues it should be entitled to First Amendment protection. For instance, if the plaintiff has direct coverage for virtual goods, it may be easier to prove the defendant’s use of the mark was “explicitly misleading” under the Rogers test. [3]

Certain platforms featuring virtual fashion items may only honor a takedown request if the complainant company has a federal registration covering goods that are the same or nearly identical to the allegedly infringing virtual goods.

The registration will provide a presumption of valid trademark rights nationwide, and it may serve as a deterrent to third parties wishing to use confusingly similar marks in virtual worlds.

B. TRADE DRESS

U.S. trademark law also protects certain source-identifying elements of a product’s aesthetic design, configuration/shape, and packaging, often referred to as “trade dress.” To obtain trade dress protection, such elements must be (1) non-functional and (2) distinctive (either inherently or acquired through use). There are a couple of interesting nuances with respect to acquiring trade dress protection in the virtual context.

First, although we have not yet seen any case law specifically addressing this, companies will likely have stronger arguments that virtual shape or design elements (as opposed to IRL elements) are non-functional. Specifically, the non-functionality requirement means the relevant elements must not be essential to the use or purpose or affect the cost or quality of the article. For real-life fashion items, this can be difficult to meet due to the inherently functional nature of many aspects of clothing or accessories. However, because virtual fashion items are essentially software code with endless possibilities, in many instances the fashion item will not require any particular design or shape to function.

Second, some virtual fashion items could receive more favorable treatment from a distinctiveness perspective. The distinctiveness requirement has historically been a difficult barrier for protecting IRL fashion. Specifically, case law prior to 2022 established that, while packaging can sometimes be inherently distinctive, product design and configuration/shape can never be, meaning companies must prove such elements have acquired distinctiveness. Proving acquired distinctiveness is burdensome because the company must have used the elements extensively, substantially exclusively, and continuously for a period of time. Often, by the time a company can acquire distinctiveness in the design, the design is no longer in style. Or, if a design is popular and copied by third parties, it can be difficult for the company to claim it used the design substantially exclusively.

If, however, a virtual fashion item provides the user with benefits that go beyond merely outfitting the avatar, such as by providing access to other products or services, one might argue that those items should be construed as packaging, or some new category of trade dress, for such other products or services, in which case the elements could possibly be deemed inherently distinctive with respect to those other products or services.

That said, if a company already has trade dress protection for IRL fashion goods, it should have good arguments that the protection extends to any virtual counterpart. On the flipside, given the difficulties companies typically face in seeking trade dress protection in IRL fashion, to the extent they can obtain trade dress protection in a virtual counterpart more easily, perhaps it can argue the rights in any virtual goods should also extend to the physical counterpart. Or, if a company introduces a physical design and virtual design simultaneously, it could possibly acquire distinctiveness in both sooner, as the simultaneous use would presumably create greater exposure to more customers and reinforce the source-identifying significance of the alleged elements.

With respect to enforcement, like traditional marks, defendants are more likely to raise a successful First Amendment defense for any virtual products allegedly infringing trade dress. The Hermès case is again an example of this, as Hermès alleged infringement of both its BIRKIN word mark and the trade dress rights in the design of its handbags, and the court held that the defendant’s MetaBirkin NFTs were entitled to the First Amendment protection.

Finally, although obtaining trade dress protection is typically more difficult than obtaining trademark protection for traditional marks such as words and logos, companies should also consider seeking registration for trade dress in virtual goods, particularly for important designs that are likely to carry over from season to season, for the same reasons discussed in the trademark section above.

C. COPYRIGHT

Copyright protects original works of authorship that contain at least a modicum of creativity, which is a relatively low bar. However, copyright does not protect useful articles. In effect, for IRL fashion items, copyright generally extends only to those designs that would be entitled to copyright protection if they were extracted or removed from the clothing or viewed on a different medium, and not to the shape of the fashion item itself.

Like trade dress protection, copyright protection should provide companies with greater protection for virtual fashion items than would be available for IRL items, particularly because the software behind the virtual fashion can theoretically create an infinite number of clothing shapes that are creative and not necessarily “useful.” Nonetheless, if a virtual clothing item is merely shaped like its IRL counterpart that lacks originality (e.g., a virtual t-shirt shaped like a basic real-life t-shirt), it may also fail to qualify for copyright protection based on a lack of creativity.

Unlike trade dress protection, however, copyright protection arises immediately upon creation of the work and its fixation in a tangible medium of expression, so it can be a useful tool for protecting virtual fashion without having to spend the time and resources required to seek registration as trade dress and establish acquired distinctiveness.

In addition, unlike IRL fashion, a separate copyright protects the underlying source code for virtual clothing items, which could provide owners with an additional, though likely limited, claim against unauthorized source code copycats.

A copyright registration will provide owners with the ability to sue for copyright infringement, but companies should balance:

  • the benefits of seeking potentially broader copyright protection in virtual fashion items (apart from the code) than it would for IRL items with the risks of conceding that virtual fashion items are works of art entitled to First Amendment protection, which would make trademark and trade dress enforcement more difficult; and
  • the benefits of obtaining any copyright registration for source code with the benefits of keeping the source code secret (although the Copyright Office permits some redactions, significant portions are required to be deposited into the public record).

We are unaware of any 2022 case law specifically addressing copyright in virtual fashion. However, the following cases are worth watching:

  • Andy Warhol Found. for Visual Arts, Inc. v. Goldsmith[4]: In October 2022, the U.S. Supreme Court heard arguments regarding whether Andy Warhol’s “Prince Series” silk screen prints and pencil drawings based on a photograph infringed the photographer’s copyright, or whether they were sufficiently “transformative” to constitute fair use. The outcome of this case could affect a copyright owner’s ability to enforce copyrights against unauthorized digital reproductions of its work, especially if the original work is fixed in a physical medium (e.g., enforcing copyright in a physical clothing item against a third party’s digital reproduction).
  • Thaler v. Perlmutter[5]: Filed in June 2022, the plaintiff is suing the U.S. Copyright Office for refusing registration of an AI-created image because there was no human author. The outcome of this case will necessarily implicate virtual fashion incorporating any AI-generated work.

D. DESIGN PATENT

Design patents protect the ornamental appearance or look of a unique product. Specifically, they protect any new, original, and ornamental design for an article of manufacture. Traditionally, this law was interpreted to require that the article of manufacture is a physical or tangible product. Thus, in the fashion industry for example, one can file a design patent application directed to a unique shoe, handbag, or jewelry design. Historically, an image or picture would not qualify for design patent protection.

However, the USPTO is currently assessing design patents with respect to new technologies such as projections, holograms, and virtual and augmented reality. In December 2020, the USPTO issued a request for public comment regarding a potential rule change to the “article of manufacture” requirement and whether U.S. law should be revised to protect digital designs. Public opinion was mixed, and in April 2022, the USPTO issued a summary of this requested information.

Although the USPTO has not yet formally revised the rules, it has issued guidelines over the years that provide examples of non-physical products that could be protected by a design patent, suggesting changes may ultimately be coming to U.S. design patent law. For example, in 1995, the USPTO released guidelines for design patent applications claiming computer-generated icons. In general, to be eligible for protection, the computer-generated icon must be embodied in a computer screen monitor, or other display monitor. The USPTO has also issued guidance allowing type font to be protectable by design patents. However, it is still unclear whether the USPTO will set forth design patent guidance specific to digital designs or virtual fashion.

Notwithstanding the possibility of obtaining a design patent specifically on such virtual goods, courts have been reluctant to find that a virtual product infringes the design patent for an IRL product. For example, in 2014, in P.S. Products, Inc. v. Activision Blizzard, Inc.,[6] P.S. Products accused Activision of infringing its design patent directed to a stun gun by depicting a virtual weapon in its video game that P.S. Products claimed resembled its patent-protected IRL product.

The court found there was no infringement because “no ordinary observer would be deceived into purchasing a video game believing it to be plaintiffs’ patented stun gun.” This case may have come out differently if the virtual gun was sold separately from the video game and could be used across various platforms rather than being one component of a particular video game. Although there are still software compatibility restrictions for virtual goods, portability of virtual goods is likely to grow as technology evolves and companies respond to consumer demands.

While we wait for further USPTO guidance that ultimately may have application to virtual fashion, parties seeking design patent protection may consider simultaneously filing one application to protect the work as a digital design on a display screen, like a patentable computer-generated icon, and a second, traditional design patent application to protect the design as a tangible product. That said, companies should consider other options for protecting any designs created by AI, as the Federal Circuit Court of Appeals held in 2022 that AI cannot qualify as an inventor for purposes of obtaining a patent.[7]

III. Virtual Fashion in Practice

Contracts relating to virtual fashion are analogous to contracts for IRL fashion and should be structured accordingly. For instance, companies should ensure that contracts with IP contributors include an assignment of all IP rights, or at least a sufficiently broad license. In the virtual context, this includes rights to the software code itself. Likewise, downstream licensing should generally address ownership, licensee rights, and if applicable, confidentiality for any trade secrets in the source code. In addition, for both IP contributors and licensees, if AI software is used in any part of the creative process, companies should give thought to allocation of ownership.

In addition, some designers or marketing teams may prefer to encourage a brand’s customer base to copy its designs or create derivative works. Although this seems counterintuitive (especially to an IP lawyer), many players in the Web3 space encourage others to build off their own designs. For example, the Bored Ape Yacht Club (BAYC), known for issuing NFTs tied to images of apes, grants owners of its NFTs the rights to use the images of apes, including for commercial purposes.[8] For example, one purchaser of a Bored Apt NFT created a Bored Ape-themed restaurant.

In the virtual fashion context, if a marketing team wants customers to build off the brand’s virtual designs but wants to retain ownership of its own designs (and perhaps derivatives), it should implement standard licensing terms relating to ownership, customer licensee rights, and other provisions. However, it’s important to consider how the terms are presented and how customers indicate assent to maximize the prospects of enforceability.

From a business perspective, companies can also now use NFTs and smart contracts to receive automatic royalties in any downstream sales or licenses. And because NFTs use blockchain technology, which provides an immutable chain of title, third parties will be able to trace such designs to the original source. This means companies can encourage the sharing of designs and receive royalties in connection with the downstream licensing of designs tied to NFTs, and third parties can confirm that the designs are legitimate by reviewing the relevant blockchain ledger. Accordingly, although encouraging customers to use the brand’s designs may not be a model for every brand, there are some steps brands can take to protect the IP rights associated with them and reap financial benefits.

As virtual fashion items become more popular, companies are faced with uncertainties and novel questions regarding how to protect and enforce their IP rights. In 2022, some questions were answered, but many more remain open. Therefore, it is important to discuss strategies for protecting innovative virtual fashion with IP counsel.

FOOTNOTES

[1] Notably, on December 30, 2022, the Hermès court denied both parties’ motions for summary judgment, with an opinion to follow by January 20. A jury trial is scheduled to begin on January 30, 2023. Hermès International, et al. v. Mason Rothschild, 1:22-cv-00384-JSR (S.D.N.Y.).

[2] See, e.g., AM Gen. LLC v. Activision Blizzard, Inc., 450 F. Supp. 3d 467, 485 (S.D.N.Y. 2020).

[3] If a defendant’s unauthorized use of a mark is protected by the First Amendment, many courts use the Rogers test to balance the plaintiff’s trademark rights with the defendant’s First Amendment right of expression. This test looks at whether the defendant’s use of the plaintiff’s mark was artistically relevant and, if so, whether it was explicitly misleading. Rogers v. Grimaldi, 875 F.2d 994 (2d Cir. 1989).

[4] 11 F.4th 26 (2d Cir. 2021), cert. granted, 142 S. Ct. 1412 (2022).

[5] Case No. 1:22-cv-01564 (D.D.C.).

[6] 140 F. Supp. 3d 795, 802 (E.D. Ark. 2014).

[7] Thaler v. Vidal, 43 F.4th 1207, 1213 (Fed. Cir. 2022).

[8] We will save for another day a discussion of the recent lawsuit against BAYC and many celebrities for failing to disclose financial incentives when promoting the BAYC NFT collection, and instead focus here on IP protection. Adonis Real, et al., v. Yuga Labs, Inc., et al., 2:22-cv-08909 (C.D. Cal.). But companies should also ensure that influencers properly disclose any incentives and other material connections.

For more intellectual property legal news, click here to visit the National Law Review.

©2023 Pierce Atwood LLP. All rights reserved.