ChatGPT: Legal implications and societal impact

ChatGPT has emerged as a groundbreaking advancement in artificial intelligence. Developed by OpenAI, it is the latest iteration of generative AI designed to output human-like text in response to prompts. It has garnered attention not only for its impressive capabilities but also for the legal and societal implications it brings to the forefront.

Legal Implications of ChatGPT

ChatGPT operates on the foundation of a deep neural network architecture. It is trained using a variety of internet-based text sources and can perform a variety of tasks such as drafting emails, answering questions, and writing code. This versatility makes it a powerful tool for a wide array of applications.

UIC Adjunct Instructor Maurine Neiberg teaches the MENG 400: Engineering Law course for the online Master of Engineering with a focus area in AI and Machine Learning program. As a practicing lawyer, Neiberg recognizes the number of legal and ethical issues associated with generative AI systems and helps break them down.

1. Copyright Infringement of ChatGPT

ChatGPT’s capability to mimic writing styles and generate text in various contexts raises questions about intellectual property. If the generative AI system creates content that is sufficiently similar to copyrighted material, we must ask who holds the rights to that content.

According to U.S. law, only humans can own a copyright. Recently, a graphic novel, Zarya of the Dawn, was under review by the U.S. Copyright Office because the author used the AI components of a drawing software Midjourney to create the book’s images. The Copyright Office ultimately decided that the individual AI-generated images would be excluded from the copyright while the authorship of the graphic novel as a whole could still be protected by copyright.

“What is generative AI and is it being used as a tool?” asks Ms. Neiberg. “If it is a tool, then the human being would be the author of the generated content and could own the rights to that work. However, if the AI system is not a tool and it’s something else, then the AI system would be the author, and what a non-human author creates would be in the public domain.”

2. Fair Use of AI systems

Fair use is a crucial concept in copyright law that allows for the limited use of material without the need for permission from the copyright holder. Engineers who create and train AI systems can use copyrighted material as training examples without the authorization of the copyright holders, and we are unsure how courts will respond to this issue. As part of the Author’s Guild, a dozen authors spearheaded by John Grisham, filed a lawsuit against ChatGPT’s parent company, OpenAI, alleging that the chatbot’s output has been built on their lives’ work. Although the case is ongoing, OpenAI claims their use of data input from the internet falls under the U.S. copyright law’s definition of fair use.

ChatGPT generates high-quality responses using existing examples that can be found online. If a user takes ChatGPT’s output and remixes it to create new work, we must ask if this is considered a transformative use or if the new content is bound by intellectual property laws.

“As an engineer, one thing to pay attention to is your proprietary knowledge,” said Dr. Carmen Lilley, Associate Professor for the UIC College of Engineering. “You are seeding proprietary knowledge because it’s being collected in a database that you have no control over. When you are thinking of your research it’s important to understand these tools and how your research is disseminated and what you are giving away.

3. Liability of using ChatGPT

Determining liability for content generated by ChatGPT is a complex issue. Is the responsibility of the AI generated content on the user providing the prompt, the developer of the model, or the model itself?

“What is the liability of a person who uses a generative AI system to produce work they’re getting paid for?” asks Neiberg. “Maybe you’ve commissioned work, but the party you hired to commission it uses a generative AI system. Or maybe you’re a publisher of a magazine and one of your paid freelancers uses a generative AI system and you publish their work. What’s your liability?”

Establishing clear lines of accountability when it comes to the content that ChatGPT creates is a challenge that requires legal frameworks to adapt and evolve.

ChatGPT’s Impact on Society

Generative AI tools are reshaping the way we work and live our everyday lives. ChatGPT has the potential to make information and services more accessible to a broader audience but has also raised concerns about job displacement and the effect it could have on our education system.

Engineers also play a crucial role in shaping the capabilities and impact of ChatGPT. Their expertise in designing the model’s architecture directly influences its performance and behavior. They are also at the forefront of addressing ethical considerations in generative AI systems.

“The technology is changing faster than the law,” said Neiberg. “These legal and ethical issues will affect engineering students because they are going to be the ones building these systems.”

The Future of Generative AI

ChatGPT represents a significant leap forward in natural language processing with a potential to revolutionize various industries. However, its implementation also introduces a range of challenges that must be addressed proactively. Collaboration between developers, engineers, and policymakers will be essential in shaping a future where AI prioritizes the well-being of our society.