IP in the age of AI: Considerations for commercial contracting with AI providers

As artificial intelligence (AI) reshapes the landscape of creation and productivity, it also raises complex questions about intellectual property (IP).  

To equip your business with the knowledge to protect your interests and thrive in the fast-evolving world of AI, here are some practical considerations around IP when entering into agreements with AI providers.

1. Training Data and Ownership 

AI systems thrive on data—vast volumes of it. Whether text, images, audio, or video, the training data that shapes an AI system’s capabilities is inherently valuable. This data constitutes IP. Therefore, understanding its origin, quality, and permissible usage is paramount.

Questions to Consider: 

• Where did the AI provider’s training data come from? 
• Is the data proprietary or sourced externally? 
• Has the AI provider obtained the necessary permissions to use third-party data?

2. The Role of Instructions (Prompts)

Instructions—the prompts given to AI systems—are often overlooked. Yet, they can significantly influence the system’s behaviour and output. These prompts can fall under IP protection and should be safeguarded to ensure they align with your strategic goals.

Questions to Ask: 

• Who owns the instructions provided to the AI system? 
• Are these instructions unique to your business processes? 
• Could they be valuable assets in their own right? 
• Can they be used by the AI provider for the benefit of third parties?  

3. The Output Dilemma

The AI-generated output—predictive analytics, creative content, or decision-making insights—is a significant area to consider. It’s the culmination of the AI’s learning process and can be immensely valuable. However, extracting value from this output requires careful consideration of IP rights.

Challenges to Address: 

•  Ownership: Who holds the IP rights to the generated content? Is it your business, the AI provider, or a combination? 
• Infringement Risks: Could the output infringe upon third-party IP?  
• Liability: AI providers often disclaim responsibility for output in their standard terms.

4. Crafting Robust Agreements

To protect your business, consider the following steps: 

• Due Diligence: Understand the AI system thoroughly. Investigate its training process, data sources, and any potential third-party entanglements. 
• Contractual Clarity: When engaging AI service providers, negotiate robust terms. Ensure that IP protections align with your business objectives and expectations. 
• Risk Mitigation: Seek support to navigate the complexities of IP law in AI agreements. Expert advice can help you avoid costly pitfalls. 

AI is a powerful ally, but it requires strategic oversight. By weaving IP considerations into your AI agreements, you can harness its potential while safeguarding your interests.

Please get in touch with our AI and technology law experts for personalised guidance.  

This article is part four in a series on AI.