Introduction
Artificial Intelligence (AI) has revolutionized the way the world works and its gaining significant popularity and investment from companies around the world. One of the most interesting applications of AI is Natural Language Processing (NPL). Natural Language Processing is a branch of AI that enables any machine to understand, interpret and generate human language as it is spoken or written. Human language is quite intricate and AI needs to be trained for it to process information which would be helpful for the machine to communicate effectively with a human. On November 30, 2022, a natural language processor called ChatGPT was introduced. It quickly gained popularity for its human-like responses and its ability to articulate information in a variety of subject areas.
As the usage of Artificial intelligence-generated content is more prevalent than ever, it is important to understand the risks involved in its usage, the AI’s limitations, as well as the ethical and legal limits that come with their use.
What is ChatGPT?
ChatGPT is a language model developed by OpenAI that has been trained using large amounts of text data and machine-learning algorithms to understand, interpret and generate human-like responses to natural language inputs. The data that is used to train ChatGPT is collected from various sources that are easily available on the internet such as articles, websites, books, and other sources. The responses that are generated by the ChatGPT model are based on the patterns and information that it has learned during its training and that training enables ChatGPT to generate human-like responses with good accuracy. The model has been trained to recognise the subtleties of the human language, such as idiomatic expressions and colloquialisms that are so unique to human communication. The GPT in ChatGPT stands for “Generative Pre-trained Transformer.” It is a type of natural language processing model that performs a wide range of natural language processing tasks including language translation, text completion and generation. ChatGPT is one of the most advanced AI systems available right now. ChatGPT surely is a remarkable invention in the tech world where the machine has had such rigorous training that it can mimic the little nuances of human-like responses but at the end of the day, it is not a human being behind the screen, writing these responses. It is a language model that only generates responses based on the patterns learned from its training. That’s why the responses generated by a model like ChatGPT might not reflect the full range of human responses. For example, a language model does not have emotions or feelings and that’s why ChatGPT lacks empathy and emotions, unlike human beings. Also, since it’s a pre-trained model, it does not always have up-to-date information. The website itself says that ChatGPT does not have accurate information after 2021 because the model has not been trained with that information yet. And lastly, it is very easy for an AI to spread misinformation or propaganda if its training data contains such misinformation or if the model doesn’t have access to accurate information. It is because ChatGPT does not have the ability to evaluate the accuracy of its responses or the information it generates. So the responsibility lies on the users to know the limitation of the AI that they use.
Copyright Implication of ChatGPT
Copyright is a type of Intellectual Property Right that gives exclusive rights to the creator of the work to control how their work is used, distributed and displayed. This means that in order to use the copyrighted work, prior permission from the creator is required. The copyright laws in India are governed by the Copyright Act of 1957. It deals with the protection of the rights of original works and their creators. Section 13 of this act lists out various categories of work in which copyright is protected. It includes:
- Original Literary work that includes books and other writings
- Original Dramatic work like plays, musicals, etc
- Original Musical work like instrumental or songs
- Cinematography films
- Original Artistic work like photography and paintings
The text generated by ChatGPT is not a direct copy of any specific text or work but the language model has been trained using a wide variety of information easily available on the internet. ChatGPT also does not hold any copyright over the content it has generated. In the end, it is still produced by a computer using pre-existing data. As a result, it is unlikely that the output of ChatGPT would be considered to be truly original and thus it is not eligible for copyright. But the data that has been used to train ChatGPT includes copyrighted material. So, the copyright of that data is still with its respective owners. Therefore, the use of ChatGPT to generate work can be a violation of copyright only if the information that ChatGPT has used infringes copyright. And if someone has heavily edited ChatGPT’s output and claims copyright over that work then they may be eligible for copyright. This is because even though the primary output was produced by ChatGPT, the final work is a result of heavy human involvement. To be considered original, a piece of work does not necessarily need to be the result of an original or inventive idea. It’s because copyright is not concerned with the originality of thoughts but with the expression of thoughts. And such expression should not be copied. This was held in the case of Camlin Pvt. Ltd. v National Pencil Industries. The copyright law protects the expression only and not the idea. Therefore, even if the idea comes from ChatGPT’s output, the expression of thought is the result of human involvement and that is protected by the Copyright Act. This includes not only text-based information but also any images or logos produced by ChatGPT because an AI does not produce these things on its own. A language model lacks the level of human creativity needed to create logos or create designs. Therefore, using that content would violate someone else’s copyright because it is unquestionably their property.
Conclusion
As an AI model, ChatGPT is a great innovation designed to understand and respond to human language. It is a valuable tool for businesses, educators, and researchers due to its incredible ability to process and analyze large amounts of information very quickly and generate relevant responses. While Artificial Intelligence offers tremendous benefits and opportunities and has the ability to ease human life it comes with its own risks and challenges. Artificial intelligence (AI) cannot think creatively or make complex decisions based on emotions and experience like humans can. And that’s why we cannot fully be dependent on AI-generated content. ChatGPT is a very sophisticated invention but it should only be used for researching or understanding something. It is designed to communicate with people in the same way that we communicate with friends or colleagues. But it is risky to blindly rely on the content produced by ChatGPT because its sources aren’t cited anywhere. And therefore it is hard to tell if a response generated by the model belongs to somebody else. Because of this, it’s crucial for copyright owners and users of these AI programmes to be informed of developments in the field of intellectual property rights.
Author(s) Name: Sanskruti Aparajit (D.E.S. Navalmal Firodia Law College, Pune)
