Generative AI offers tremendous opportunities for businesses to fast-track development, streamline knowledge synthesis, and sidestep mundane tasks. However, when deliberating agreements with generative AI providers, it’s crucial to ensure protection of business interests, intellectual property, and confidential data under US law. For some examples of how AI can apply in corporate law, check out our previous article.
Defining Generative AI: Generative AI refers to AI technologies that discern patterns from training data and utilize this knowledge on user inputs, generating fresh content, from text to multimedia.
Generative AI promises numerous benefits, but it’s not without risks. For a comprehensive overview of these aspects, refer to our guide: “Top Ten Considerations for Companies Using Generative AI Internally.”
Note: While this article offers insights on generative AI usage terms, it doesn’t replace legal advice. It’s a broad guide, not an exhaustive review.
Understanding AI Agreements: Generally, using generative AI necessitates an agreement with the tool provider. These can be online terms of service, bespoke agreements, or terms of use. They outline ownership rights, confidentiality clauses, and other legal stipulations.
Providers may offer “business” versions with more favorable terms than the generic user agreement. You might also be able to tailor an agreement to your firm’s needs. However, superior terms might come with additional costs.
Moreover, some providers offer varied service types, like plugins or APIs, which might have distinct terms from the basic tool.
Prompt and Output Rights:
- Prompts: Often, the user retains ownership of the prompt. However, providers may reserve extensive rights to utilize and even redistribute the prompt. Using valuable proprietary data in prompts can indirectly benefit others, possibly including competitors.
- Outputs: Users typically own generated outputs. Still, terms may grant providers rights to the content, potentially for AI training or distribution. If this is a concern, scrutinize available account types and terms, or rethink your usage strategy.
Addressing Confidentiality: Most terms of use don’t impose confidentiality on the provider. Without these obligations, there’s no contractual breach recourse if the provider discloses information. Sharing third-party confidential data or personal details can also lead to legal issues. An internal generative AI policy can mitigate risks tied to handling sensitive information.
Indemnification Stipulations: A majority of providers don’t offer indemnification to users, but exceptions might exist for business accounts. Contrastingly, users might be bound by broad indemnity clauses for any liabilities arising from tool usage. Consequently, using generative AI could expose businesses to considerable indemnification responsibilities.
Liability Limitations: Frequently, providers minimize their contractual liabilities. Standard online agreements might absolve them from certain damages, with some even capping liability for direct damages. Conversely, users rarely benefit from significant liability limitations. Thus, without a negotiated agreement, liability risks might heavily lean against the user.
Final Thoughts: Generative AI is a game-changer, but it’s pivotal to understand the legal and contractual implications. We urge businesses to holistically evaluate the pros, cons, and terms of generative AI tools under consideration, ensuring an informed and balanced decision.
