With the use of non-human intelligence, generative artificial intelligence, or GenAI, is able to produce and control human-like concepts and thought processes. These programs, which include Google’s BARD and OpenAI’s ChatGPT/GPT-4, among others, are first trained on big data sets and computer power. Following the instruction, people can improve themselves to provide original, tailored material. Given that content generation, which was previously dependent on human inputs, has progressed outside that domain, this has presented new problems to copyright experts. Now, the user receives personalized information instead of replies derived from their search queries, as found through Google. The use of GenAI in the creation of this new content has raised questions about copyright infringement, privacy violations, libel, and defamation, among other issues. Because corporations are using user-provided data, which includes data generated by minors, to train these programs, copyright infringement is especially concerning because it increases the vulnerability of these programs. There are concerns about the extent to which businesses can rely on the Copyright Act’s “fair use” exception.
Gen-AI tools and unprecedented levels of copyright infringement
The new oil of the twenty-first century is data. Everybody generates data in their daily lives, and AI developers have frequently come under fire for misusing it. First off, the AI developers have come under fire for allegedly “web scraping,” or stealing user data without permission. Over 300 billion words from the internet, including “articles, websites, books, and posts, including personal information obtained without users’ consent,” are allegedly stolen by OpenAI in order to train their software. Since data is now regarded by US courts as “property,” scarping gives rise to claims of theft and misuse of data.
Second, we still have a legitimate expectation that our data will be safe and secure, even in the case that we publish a photograph or any other kind of data online for public viewing, such as on our blog or social media profile. The Supreme Court acknowledged the right to privacy as a component of Article 21 of the Constitution in the case of KS Puttaswamy v. Union of India. When our data is utilized to train AI software without our express authorization, this careless behavior violates our privacy and creates a breach of trust.
Thirdly, children’s data and privacy issues are placed on a different level. They are more vulnerable to mistreatment, prejudice, and exploitation. Regretfully, AI doesn’t have enough tools to stop kids from using its portal. The Children’s Online Privacy Protection Act (COPPA) in the USA prohibits tracking, gathering, or utilizing data about minors unless legitimate parental or guardian approval is secured. Sadly, India lacks these regulations; the only available check is self-reporting, which is ineffective.
Fourthly, when asked to credit a specific author’s work, such as a poem, AI software has also generated erroneous or fraudulent paragraphs. Such an inaccurate outcome misleads the audience and violates the writers’ moral rights by misrepresenting their works. The next section will discuss the ways in which the courts are handling these issues, providing guidance for the future development of copyright jurisprudence.
Using Judicial intelligence to solve the generative Artificial Intelligence Issue
“Learning from machines is not a reason to breach the law. The information utilized to enhance algorithms must be legitimately gathered and stored. According to OpenAI, the fair-use exception applies to the data used for AI training and execution. A few examples of fair usage exceptions are outlined in Section 52 of India’s Copyright Act, 1957. These include, but are not limited to, “reading or reciting in public reasonable extracts from a published literary or dramatic work” and “the reproduction for the purpose of research.”
In order to assess whether there has been copyright infringement in the outcome of the user’s query, US courts now utilize a two-pronged approach. The plaintiff must demonstrate two things: first, that the program actually copied the underlying work by having “access to their works”; and second, that the software produced an output that was “substantially” identical. The second criterion in this case is hard to pin down because it takes into account a lot of other things, such as “similar concept and feel” or “overall look and feel” or the fact that a normal, reasonable person couldn’t tell the two pieces apart. This is a judgment call made for each result the AI program generates in answer to a user’s inquiry.
This implies that the AI user, whose search query resulted in information that violates copyright, and the AI corporation, which offered the platform, may both be held accountable for copyright infringement. “Generally, they do not regenerate unaltered data from any particular work in their training corpus,” according to OpenAI, their AI systems claim. Even so, if OpenAI was unable to stop an infringement, it might still be held accountable under the legal doctrine of “vicarious liability.”
Moving on, the following part clarifies the evolving Indian jurisprudence while addressing the questions surrounding the authorship of texts produced by AI.
Indian Jurisprudence on Gen AI: Historical, Current, and Prospective
Only works that are “created by humans” are recognized as being eligible for protection by the US Copyright Office. The original owner of the work is the one who created it; non-human authors are not protected by copyright. In general, things proceed similarly throughout India. Any “person” who initiates the creation of a computer-generated work is referred to as an “author” under Section 2(d) of the Copyright Act, 1957. Such work must be “original,” as required by Section 13, yet the statute gives no definition for the term. Regarding granting protection to content created by AI, the Indian Copyright Office has shown hesitancy. In the past, when AI co-wrote a work with a human author, it issued a withdrawal notice.
The Copyright Act, 1957, is “not well equipped to facilitate authorship and ownership by artificial intelligence,” according to the 161st Parliamentary Standing Committee Report. For example, Section 16 of the Act states unequivocally that “no person” shall be entitled to copyright protection except in compliance with this Act’s provisions. To “incentivize innovation and R&D,” the report recommended that “patent protection” be extended to works created by AI. To foster innovation and improve AI-based expressions, we propose that this logic be expanded to include copyright protection for content created by AI.
In addition, the “right to be forgotten” is established under Section 12 of India’s newly passed Digital Personal Data Protection Act, 2023, which was enacted on August 11, 2023. It requires that the user’s personal information be deleted at his or her request. This could serve as a simple method for getting rid of content that violates copyright. Practical issues occur, though, because AI software cannot “unlearn” after it has been taught on a dataset.
The exploitation of AI tools for copyright infringement has been actively restrained by Indian courts. The court in Anil Kapoor v. Simply Life India imposed an order prohibiting the use of AI to produce phony or altered information, particularly for profit. Its goal was to defend each person’s right to their own unique identity. The US District Court ruled in Mareta v. Google Inc. that remedial measures, such as those intended to prevent copyright infringement and safeguard privacy, must be interpreted broadly to encompass “new technologies.” The Indian courts have also acknowledged this fundamental notion of fast change. Nonetheless, “there is a need to review the provisions of [the Copyright Act] on a priority basis,” as the 161st Parliamentary Report recommended.
Concluding remarks and suggestions
Artificial intelligence (AI)-Generated content is here to stay and is becoming more and more integrated into our daily lives. This makes it imperative to have in-depth conversations about the elephant in the room. The 161st Parliamentary Report’s recommendations to “create a separate category of rights for AI and AI-related works” are once again supported by us. In light of Section 12 of the Digital Personal Data Protection Act, 2023, there is a need for efficient transparency and cybersecurity processes, including data disclosure and guaranteeing the right to erasure. Secondly, to establish an AI Fund to address any future violations immediately, as well as to guarantee legal provisions to stop the improper use of children’s data and safeguard the right to privacy granted under Article 21.
Adv. Khanak Sharma (D/1710/2023)