AI case by ‘NYT’ a guardrail for journalists’ copyrighted works

There is a new technological wave known as generative artificial intelligence (generative AI) pioneered in November last year by an AI lab, Open AI. The chatbot eerily articulated and nuanced text in response to short prompts with people using it to write love letters, poetry, fan fiction and school work. AI breakthroughs promise an unprecedented opportunity but also economic, political and social disruption.

In the face of such tech advancement, there is, indeed, cause for both concern and optimism. This is evident in a case currently sub judice before a New York District court, where The New York Times (NYT) has sued Open AI and Microsoft for copyright infringement over several news articles, some even lifted word for word, and most data extracted for free from the internet scrapping. This data was trained to the duo’s large language models (LLMs) and accessible to their users commercially.

However, these AI companies have made several rebuttals, arguing fair usage of copyrighted content for a new “transformative purpose”. “Legitimate interest” has also been argued forth. All these fly in the face of “limited use” with “consent” of the original authors. Even here in Kenya, several works, including from indigenous language authors, was indiscriminately taken and trained to LLMs with neither consent of nor compensation to the intellectual property owners. And without any colour of right.

With this case in court, journalists can feel safe in the advent of the looming generative AI threat that their copyrighted works are legally protected and any attempt at illegal use of their work shall clearly invite legal repercussions. Intellectual Property is protected by copyright laws and journalists’ copyrighted works are protected; only used through consent unless limited use for fair usage to qualify.

Fair use and fair dealing are exceptions applicable to the use of copyrighted works. They allow third-party use of copyrighted work without the owner’s permission but not raising copyright infringement claims. The Second Schedule of the Copyright Amendment Act, 2019 allows for fair dealing (What amounts to fair dealing in Kenya?) It groups the exceptions into general exceptions and limitations like scientific research, private use or use by a single person, criticism or review or the reporting of current events; educational institutions; libraries and archives; and broadcasting.

The limitation of the rights arising from copyright protection allows for just, equitable and flexible use of authors’ works. Accordingly, acknowledging and attributing the authorship of one’s work to the rightful owner is part of fair use of the work. Therefore, the fair use of copyrighted works should always be done in good faith and, where there is monetary gain as a result of such use, fair compensation made to the author of the respective work.

A US court in Washington, DC, ruled that a work of art created by AI without human input cannot be copyrighted under US law, says an August 21, 2023 Reuters article by Blake Brittain, “AI-generated art cannot receive copyrights, US court says”. The court further ruled that only works with human authors can receive copyrights, per US District Judge Beryl Howell, affirming the Copyright Office’s rejection of an application filed by computer scientist Stephen Thaler on behalf of his DABUS system.

Beware, AI is an ever-looming threat to copyrighted works. AI regulation and supervision is timely, to ebb the trend of copyright infringement. At least from the ongoing case, the damages can go a long way to compensate journalists.

Mr Ayuo, an Alan Turing Ethical AI Fellow, is a legal researcher. [email protected].