Regardless of if you find yourself in the business world, academia, or simply browsing the Internet, it is nearly impossible to escape the discourse surrounding the recent insurgence of artificial intelligence. Its capabilities appear to be boundless, and many individuals have chosen to utilize its generative abilities to create art, brainstorm business ideas, and problem solve real world issues. However, it is incredibly important to note—especially for entrepreneurs and business owners just setting out—that there are potential pitfalls hidden in the glamour and ease of artificial intelligence that must be carefully navigated.
Protecting AI-generated work
Building an intellectual property portfolio is often on the first page of an entrepreneur’s to-do list: it is imperative to ensure that a business’s name, slogan, and logo are secured by trademarks, and that patents are procured for any inventions relevant to the company’s product or service offerings. Business owners must remember that as tempting as it may be to have artificial intelligence help out with designing the company’s branding or generating an eye-catching logo, the resulting images would not be protectable under any intellectual property options, a fact that is not emphasized by AI developers who may be pushing their software as mechanisms for services like logo creation. This is because material generated by AI lacks a human creator, which is a key factor that must be considered when obtaining a patent or copyright registration.
In 2022, the Thaler v. Vidal case addressed whether or not food containers that were created without any human contribution were patentable. The United States Patent and Trademark Office had originally rejected the patent applications based on their lack of human input, and the U.S. Court of Appeals for the Federal Circuit agreed, stating that statutory interpretation limits inventorship to natural persons, which obviously disqualifies AI mechanisms—but what about situations in which the subject matter in question was only partially created by AI? The U.S. Copyright Office considered this scenario when Kris Kashtanova attempted to obtain copyright protection for a partially AI-generated comic book. The images within the comics were AI-generated but human-selected, and Kashtanova wrote the text and oversaw the final arrangement of the images and overall book. The USCO ultimately concluded that while the text and arrangement of it were sufficient for a copyright registration, they made sure to specify that the registration excluded any artwork generated by artificial intelligence.
While many entrepreneurs may just be using AI to generate images for a company logo and not necessarily an entire comic book, it is still imperative to understand the implications of these decisions: a logo that may have been otherwise able to obtain a trademark or copyright registration is disqualified if created with AI. That is not to say that such a logo would be impossible to use, but as more and more businesses turn to AI for these creative needs, there may be an influx of highly similar branding onto overlapping markets, resulting in consumer confusion between product and service source identifiers—which is already one of the biggest issues intellectual property registrations strive to combat.
Safeguarding art against AI use
On the flip side of this issue, creators are discovering that their artwork can be used as training data for artificial intelligence platforms. Generative AI ‘learns’ to create by looking at pre-existing work and then adjusting boundaries accordingly for queries by processing massive databases of text and images. This information is then used to create rules for the software, allowing it to respond appropriately when prompted. This suggests that some of the work that is flowing through this mechanism can be and likely is subject to some kind of intellectual property protection, carrying with it major legal implications. To avoid infringement upon another party’s intellectual property, some companies have gone the route of training their AI models only with content in which they already owned rights, limiting potential future conflict.
In Andersen v. Stability AI et al., three artists formed a class action suit against several AI platforms, alleging that their original work was being used to train AI in their specific artistic style, resulting in the capability of generating works riffing off of their own without license. While the case is still pending, the U.S. District Court for the Northern District of California is likely to find that images altered by generative AI are not automatically in violation of the Copyright Act as derivative works; this is not entirely unexpected, as rulings like this one were suspected to lean on interpretations of the fair use doctrine, which allows copyrighted work to be utilized without permission for various uses, including a transformative manner.
While this arena still maintains an air of uncertainty, business owners and entrepreneurs looking to protect their intellectual property are encouraged to do so in an effort to mitigate risk and create a long-term protection plan. Individuals creating content or looking to safeguard their branding should police the marketplace to keep tabs on any potential infringers. At this time, responsibly maintaining trademark and copyright registrations is the best way to remain prepared to deal with infringers, whether they be AI-generated or not. Businesses undertaking transactions with AI platforms especially should require proper licensure for data, training or otherwise, as well as add disclosures for customers and partners alike regarding any use of AI.
Overall, it is unavoidable that artificial intelligence is here to stay; the path forward is best constructed by establishing how to best work with and around it, ensuring that all parties’ rights are preserved and creativity is respected, all while enjoying the distinct advantages that come with utilizing such unprecedented and far-reaching software. Advocates, creatives, and entrepreneurs alike must adjust to new industry realities facilitated by AI, and one of the best ways to do so is by establishing a robust intellectual property portfolio.
https://hbr.org/2023/04/generative-ai-has-an-intellectual-property-problem
https://www.msk.com/newsroom-alerts-Federal-Judge-Dismissive-of-AI-Complaint
Images: https://www.freepik.com/free-photos-vectors/artificial-intelligence-illustration