Introduction

Generative AI, a technology that harnesses massive bodies of data to create art, text, and other content, is rapidly transforming creative industries. While it introduces a world of possibilities, it also brings a plethora of legal issues, particularly regarding copyright infringement. In this blog post, we will delve into the legal challenges associated with AI image generation, which often involves the unauthorized use of artists' work and models' likenesses. Finally, we will also discuss Adobe and Nvidia's innovative approach to mitigate these legal concerns.

Generative AI and Clouds of Doubt Surrounding its Usage

Generative AI platforms, like Stable Diffusion, Midjourney, or DALL·E 2, create striking visuals and text by processing vast amounts of data. However, these powerful AI tools raise important legal questions that are still in the process of being addressed. Moreover, here are some of the key legal issues:

1. Intellectual Property Infringement: Generative AI's ability to generate content raises questions about intellectual property rights. This includes copyright, patent, and trademark infringement.

2. Ownership of AI-Generated Works: Who owns the content created by generative AI platforms – the AI developer, the user, or the referenced work of the content creator?

3. Unlicensed Content in Training Data: Many generative AI models are trained on vast datasets. These datasets may contain unlicensed or copyrighted works, leading to potential infringement claims.

4. Fair Use Doctrine: The legal system challanges to clarify what constitutes a "derivative work" under intellectual property laws. However, this leads to gray area with interpretations varying by jurisdiction.

5. Transformative Use: An example of a case cases is the Andy Warhol Foundation lawsuit. It is filed against photographer Lynn Goldsmith. This case will shape how AI-generated works are treated and whether they are considered "transformative."

Legal Cases and Implications

However, several legal cases have emerged that highlight the challenges of generative AI. For example, in the case of Andersen v. Stability AI et al., artists sued generative AI platforms for using their original works without permission to train their AI, leading to the creation of potentially unauthorized derivative works.

Another lawsuit saw Getty Images suing Stable Diffusion for allegedly using its photos without proper licensing, violating copyright and trademark rights. These legal battles are forcing the legal system to clarify the boundaries of derivative works under intellectual property laws.

Mitigating Legal Risks and Building a Way Forward

In light of these legal challenges, various companies, stakeholders, and even individual creatives must take steps to protect themselves and promote responsible AI usage:

AI Developers:

• Ensure compliance with the law regarding data acquisition and licensing.
• Compensate individuals who own the IP used in training data.
• Provide transparency on the origin of data used in AI models.

Content Creators:

• Actively monitor data lakes and datasets for unauthorized use of their content.
• Utilize automation tools to search for copyrighted elements in large datasets.
• Consider co-creating with followers, ensuring they give permission for content usage.

Businesses:

• Demand terms of service that confirm proper licensing of training data from generative AI providers.
• Add AI-related language to confidentiality provisions in contracts.
• Evaluate transaction terms to protect intellectual property rights.

Adobe and Nvidia's Innovative Approach

Firstly, Adobe introduced "Firefly," an AI-enhanced feature that allows users to generate images, illustrations, or videos using words. Furthermore, Adobe updated Firefly in October 2023, vastly improving it's photorealism. On top of that, Adobe added image generators to nearly all of their creative cloud including vector AI in Adobe Illustrator and text-to-image generators in photoshop. Additionally, these creations are based on Adobe stock images, publicly licensed content, and content with expired copyrights, making them safe for commercial use. Lastly, Adobe advocates for a "no training" label, enabling content creators to request that their content not be used to train AI models.

Nvidia, on the other hand, launched "Picasso AI Image Generator," which pays royalties on images licensed by Getty Images, Shutterstock Inc., and Adobe. These initiatives represent a significant step in reconciling copyright and payment issues in generative AI technology.

Conclusion

While generative AI opens up new creative possibilities, it also introduces complex legal challenges, particularly in the realm of copyright and intellectual property. Companies and content creators must proactively protect their interests and demand transparency from AI providers. Adobe and Nvidia's deserve credit in progressing the usability of AI generated content. Finally, their innovative solutions aim to bridge the gap between AI technology and legal compliance, offering a path forward in the evolving landscape of generative AI.