Two Facets of AI and Law: Creativity and Privacy
“Creativity is the process by which the human mind shines out a beam of light in the darkness to assert its uniqueness in an increasingly complex and crowded world…. Privacy is that space we retreat to rediscover and safeguard ourselves from the outside world”.
Starting the discussion with this explanation of what the two elements he was touching upon meant, Mr. Srinath proceeded to explain why he had picked creativity and privacy as the foundation of the discussion; both are essential for one to achieve self-actualisation. He emphasized that automation has always had the effect of replacing humans in some form or the other, but he asserts that with AI, this is different: none of these previous technologies even had the potential for sentience.
Moving on to creativity, Mr. Srinath picked famous pieces of art that were made with the help of AI generation tools to explain the United States’ position on providing copyright protection to AI generated content. Primarily drawing on James Allen’s “Théâtre d’Opéra Spatial” and Stephen Thaler’s “A Recent Entrance to Paradise”, he pointed out a subtle distinction in the two decisions. The former looked at the predominance of a creative human element over AI-generated work, while in the Thaler decision, ownership was required to be vested in a human for copyright protection. In both cases, copyright protection was denied. He proceeded to concisely state the U.S. position: a human can be said to own a copyright only when they exercise the power of creative thought in the creation of an artwork. He then dealt with the reasoning in the case where, for the first time, the U.S. Copyright Office granted copyright protection to an artwork involving AI-Generation – “A Single Piece of American Cheese” by Kent Keirsey. He emphasised the reason for the grant of copyright – there was sufficient creative input from the artist, and the creative elements of the art piece were more than the AI-generated elements.
He then explained the United Kingdom’s position on Stephen Thaler’s patent application, where again, there was a need for a human author for IPR protection. Moving on to the Indian position, to obtain copyright protection, he emphasized that there must be a work that can be copyrighted, and the work must have an author. Drawing on more comparative jurisprudence, he concluded that the trend across the world seemed to be a refusal of copyright for AI-generated content due to the lack of the human creative element. But he also mentioned some exceptions where AI-generated content was awarded IPR protection, like South Africa.
By further playing out interesting AI-generated videos that were made out of iconic paintings, he emphasized that the creative element was really lacking in them. He asserted that in such cases, the human element of entering a prompt was very minimal. Thus, Mr. Srinath posited that even if AI can generate novel content, the law will deny copyright protection unless it is shown that the “creative genius” of the human mind was instrumental in the creation of the work. Creativity formed the base of ownership of the content for him. He further touched upon the understanding of property drawing from the interplay between copyright and AI in this context.
Next followed an interesting element of the talk, where he required four volunteers for an experiment, aiming to demonstrate the difference between human and machine creativity. This experiment was based on the understanding of cognitive schema, that is, the way in which our brain absorbs content using our senses, and retention is based on past life experiences. He explained that when a human receives a certain input, and converts that into a creative output, it is by their cognition that the data is processed, assimilated, and sent out as creative output. This was what was accorded legal protection under copyright law in his view. For the experiment, he asked students to sequentially add an element of creativity and recite a famous legal quote: “We have buried the forms of action, and yet they rule us from their graves.” Each student added a creative element to the line, which was to be imitated by the next student with an addition of their own creative spin over the line. Using this experiment, he highlighted how with each additional human in the experiment, the creative element in delivering the line increased. He then referenced an example of ChatGPT being asked to imitate a person’s photograph as similarly as possible, one hundred times. The result was that over the imitations, the person’s face became increasingly distorted. Comparing this example with the experiment that had just been made in the lecture, he asserted that “AI’s creativity over iterations drifts downward. Human creativity over iterations drifts upwards.”
Following this, he proceeded with the discussion on Privacy. He explained the right to be forgotten recognised by Indian Courts, drawing from the case of Karthick Theodore v. Registrar General (2024). He stated that AI-based algorithms work on large volumes of data that include everything from our pasts, stripping us of our privacy. For instance, what would happen if all of a person’s anti-government posts online were to be tracked using AI, and this data were used to deny a constitutional post or a government job? But then he also drew on examples where a trade-off between an intrusion of privacy was countered by a greater good to society, and posed a question to the attendees as to whether that trade-off is acceptable. He then explained the “Nothing to fear, Nothing to Hide” fallacy. He posited that one must never surrender our rights to any government, simply because they say that if one has nothing to hide, then there should be no problem in showing their data or actions. The right to privacy is enshrined in the Constitution, regardless of there being nothing to hide.
Mr. Srinath then laid down a framework of the Indian law on data protection and privacy, explaining the difference between a data fiduciary and a data processor under the Digital Personal Data Protection Act, 2023. The data fiduciary is the one who collects the data, and it is given to the data processor to process the same. There is an obligation on the data fiduciary to delete the data upon the cessation of the service relationship. However, in cases where the data fiduciary provides the data to be processed to an AI algorithm, he posed the following question: Is AI algorithm the data fiduciary or the data processor? He drew attention to the criminal penalties applicable in case of non-compliance with the obligations under the DPDP Act too. He pointed out a peculiar aspect with which we treat AI today: anthropomorphism. This refers to the growing inclination to ascribe a human form to a non-living entity. Finally, he touched upon the black-box paradigm – the inability to understand or know the reasoning using which an AI algorithm to come to a particular conclusion in the form of an output. He asserted that we are comfortable with the “human black boxes”, but may not have the same level of comfort with an AI black box.
The discussion concluded with a short round of questions from the audience, and a vote of thanks from the organisers.