Trade Secrets Emerge as Key Strategy for Protecting AI-Generated Intellectual Property Amid Patent and Copyright Challenges

Artificial intelligence (AI) has become ubiquitous across virtually every industry, prompting critical questions about the scope and applicability of intellectual property (IP) law to AI-generated creations. As AI tools like ChatGPT and DALL-E gain prominence, companies must navigate uncertain legal terrain concerning ownership, protection, and liability related to AI outputs and platforms.

AI broadly refers to computer systems capable of performing tasks typically requiring human intelligence. Generative AI, a subset of AI, creates novel content by analyzing vast datasets through deep learning techniques. This capability raises unique IP challenges because the outputs are original rather than mere reproductions of existing works.

Despite AI's growing commercial impact, legislative and judicial responses remain limited. Courts have yet to establish definitive precedents, though several high-profile cases are emerging that may shape future IP law concerning AI.

One pressing concern for technology companies is how to protect their AI platforms and proprietary tools. Patents, traditionally a strong form of IP protection, present significant obstacles for AI-related inventions. Under the Supreme Court’s Alice framework, abstract ideas—including many software algorithms—are ineligible for patent protection. AI systems often function as “black boxes,” making it difficult to clearly define inventive steps beyond abstract concepts or efficiency improvements.

In contrast, trade secret law offers a promising alternative. The federal Defend Trade Secrets Act protects confidential business, scientific, and technical information that derives economic value from its secrecy, provided reasonable measures are taken to maintain confidentiality. This framework can encompass AI training data, source code, input parameters, and internally used AI-generated outputs. Unlike patents, trade secrets can potentially last indefinitely, though they require ongoing efforts to preserve secrecy.

However, once AI-generated outputs become public, they lose trade secret protection. This limitation is particularly relevant given the increasing public dissemination of AI-created works.

Protecting AI-generated content through copyright and patent law remains challenging. Courts and copyright offices generally require human authorship or inventorship. For instance, Stephen Thaler’s AI platform DABUS has been denied patent recognition as an inventor in the United States, where the law mandates inventors be natural persons. Some jurisdictions, such as South Africa, have diverged by granting patents to AI-generated inventions, though rulings vary globally.

Similarly, copyright protection for AI-created works is typically denied absent human creative input. Thaler’s “Creativity Machine” artwork was refused copyright registration, with courts affirming that U.S. copyright law protects only human-created works. When humans collaborate with AI, copyright protection may extend only to the human-contributed elements.

Trade secret law circumvents the human authorship requirement, focusing instead on the secrecy and economic value of the information. Consequently, AI-generated works kept confidential may qualify as trade secrets, offering a practical protection mechanism for companies seeking to safeguard proprietary AI outputs.

In summary, while patent and copyright protections for AI-generated intellectual property face significant legal barriers, trade secret law provides a flexible and durable alternative. Companies leveraging AI technologies should prioritize robust data protection, carefully document human contributions to AI outputs, and monitor evolving legal developments across jurisdictions to mitigate risks and maximize IP value.