Microsoft recently announced a large investment in OpenAI, a company that created ChatGPT, an artificial intelligence (AI) platform. The amount Microsoft invested is yet to be disclosed but it is rumored that they invested approximately $10 billion for 49% of the company.
Venture-capital firms’ fundraising hit a nine-year low, raising $20.6 billion in the fourth quarter; this represents a 65% drop from the previous year and is the lowest amount raised in the fourth quarter since 2013. While the venture-capital market is mostly in decline, AI companies remain a bright spot. Erin Prince-Wright, a partner at Index Ventures, expects the rate of venture capital dollars going into AI to continue to accelerate, and predicts AI could become an essential technology used in almost all contemporary businesses. This helps to explain why Microsoft was willing to invest such a large amount in OpenAI despite the rising prevalence of AI, but what is ChatGPT?
What is ChatGPT?
ChatGPT is a free online resource that users can prompt with plain language inputs and responds to questions using artificial intelligence (AI.) A few of the examples the AI’s functionality include producing written content such as essays, stories, or articles, writing programming code, and solving math problems. The tool reached one million users just five days after it was launched in November 2022. This is especially staggering considering it took Spotify and Netflix five months and three years respectively to reach the same million-user milestone.
Professors from the University of Minnesota’s Law School and the University of Pennsylvania’s Wharton School of Business recently put ChatGPT’s abilities to the test. The AI passed four law school exams averaging a C+ grade and also earned a B grade on a business management paper.
The New York City Department of Education and Seattle Public School System have banned students from using ChatGPT due to concerns about plagiarism and cheating. OpenAI is developing ways for schools to detect AI plagiarism but the CEO compares ChatGPT to tools like calculators and says that schools will need to adapt to generative text as determined individuals will find ways to avoid detection.
Copyright
Another important legal consideration is how copyright law and questions of ownership will impact works produced by ChatGPT. Copyright is a form of intellectual property intended to protect original works of authorship that the author fixes in a tangible medium of expression, including works like books, movies, paintings, and even blog posts. Works are deemed to be original when they possess a minimal degree of creativity and are created independently by a human author. ChatGPT is capable of generating creative and original textual works, which begs the question, who if anyone owns the copyright to ChatGPT’s creations? Some may want to credit the program itself, or perhaps the user of the program providing inputs into it can be said to be the author of the work.
Who Owns Copyright to ChatGPT Work Product?
The AI?
While U.S. copyright law doesn’t explicitly state rules for non-humans, courts have consistently found “that non-human expression is ineligible for copyright protection.” Notably, the People for the Ethical Treatment of Animals (PETA) filed a lawsuit against a photographer, David Slater, on behalf of Naruto, a crested macaque (primate.)
Naruto took multiple selfies using Slater’s camera, and Slater sold a book including some of the photos. The court found that Naruto lacked standing to sue under the Copyright Act because the act uses “[t]he terms ‘children,’ ‘grandchildren,’ ‘legitimate,’ ‘widow,’ and ‘widower’ [which] all imply humanity and necessarily exclude animals, of which do not marry and do not have heirs entitled to property by law.
Faced with the question of whether an AI can hold copyright to the work product it creates, a court today would likely follow this logic and precedent and come to the conclusion that AI cannot hold copyright as it lacks humanity.
The AI’s Creator?
The AI’s creator also seems to be ineligible for copyright protection for work created by the AI. Stephen Thaler created an algorithm called the Creativity Machine designed to repurpose images to create what is seen by a dying brain. Thaler sought copyright protection for images created by his AI but the USCO found that “human authorship” was lacking and denied him copyright protection.
The user who provides the prompt?
The user who provides the prompt to the AI seems to be ineligible for copyright protection as well. Kris Kashtanova created a comic book using images created by a text-to-image AI, Midjourney. She was initially granted copyright protection by the United States Copyright Office (USCO). However, she did not disclose that the illustrations were generated by an AI.
Once the USCO learned that the images in the comic book were AI generated, they informed Ms. Kashtanova that they intend to cancel the original certificate and grant her a new one that only covers the elements of the work she created. The USCO determined that Kashtanova is the author of the comic book’s text along with the arrangement and selection of visual elements, and that these elements of the work are eligible for copyright protection.
Kashtanova considers this result “great news” but is considering how to best argue that she was the author of the images as well. The general counsel who represents Midjourney is optimistic and understands the USCO’s characterization as, “if an artist exerts creative control over an image generating tool… the output is protectable.” This decision regarding AI generated content provides some much needed clarity but the question remains: how will copyright law evolve to accommodate more sophisticated AI in the future?
Conclusion
Regulation on AI-generated content is still developing. Here’s what we know so far:
1. Copyright protection is a manifestation of the Constitutional provision Article 1, Section 8, Clause 8:
“To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries”
2. The Copyright Act 17 U.S.C. § 102(a) only protects works that are original, which requires (1) independent, human authorship and (2) sufficient creativity.
Since one of the main reasons for intellectual property rights is to promote human creativity, it follows that artwork authored by nonhuman entities that fail to require even a modicum of human creativity would not be eligible for protection.
Based on items 1 and 2, the U.S. Copyright Office determined that the images in Kashtanova’s comic book were not eligible for copyright protection, even those that had been edited in Photoshop by Kashtanova. However, the text of the comic book is eligible for protection because Kashtanova wrote the story without the help of any tools. Fortunately for Kashtanova, the selection and arrangement of the images itself is eligible for copyright protection.
The U.S. Copyright Office’s decision still leaves some AI-generated content in question. For instance, if the tempo, melody, form, and key were sourced from an AI generating tool could the arrangement of those elements as a song be eligible for copyright? Although AI-generated content is the center of attention today, its impact on art and business as we know it will be determined by the decisions on its eligibility for legal protection.
*The views expressed in this article do not represent the views of Santa Clara University.
Comments