As AI continues to evolve and become integral across industries, including in the marketing and digital advertising space, it is important to not only address the use of AI in your Master Services Agreement (MSA) or other services contract with your clients, vendors, and employees, but to also understand the terms and conditions of the AI platforms you use. These terms and conditions play a crucial role in defining the relationship between users and the AI platform.

Please note that AI is constantly evolving, which means the terms and conditions are likely to continue to evolve too. We recommend you regularly check AI platforms you use for updates to their terms.

In this blog, we provide a summary of the key terms and conditions of the biggest AI platforms, with a particular focus on comparing indemnification and liability clauses.

*NOTE: It’s important to note the time stamp here, which is late March of 2025. We’ll do our best to update this on a consistent basis, but be cautious if you’re seeing this information significantly later than that.

Terms of Some Popular or Dominant Platforms

Platform: Google (Gemini)

Google retains rights in its content but does not claim rights in any original content you create. However, by using Gemini, you acknowledge that Google may produce similar content for other users.

Google distinguishes between paid and unpaid services. For unpaid services, Google can use the content you submit to further develop Google technologies. For paid services, Google does not use your prompts or responses to improve its technologies but does process or log them for monitoring purposes.

By using Gemini, you agree that the technology is experimental, you must use discretion, and you should not rely on the content for professional advice.

Google has a specific Generative AI prohibited use policy that is standard and prohibits dangerous, illegal, discriminatory, or misleading activity.

Google’s general indemnification provision stipulates that users indemnify Google and its affiliates for any claims arising from their violation of the terms of service. For Google’s AI platforms and their specific AI policies, including Gemini, Google Cloud, and other services, indemnification provisions may vary. Google’s general terms also specify that Google may be held liable for its breaches of the terms of use, and there is a limit to their liability, except in instances of gross negligence or willful misconduct.

Platform: ChatGPT (Open AI)

Open AI’s terms include standard language that you cannot use the platform for illegal or abusive activity, that it is not responsible for any third-party sources or software, that you as the user own, or have secured the rights to, all input and own all output. You also acknowledge that output content may be similar to that of other users.

Open AI may use input or output content to improve its services, unless you choose to opt out. There is also a limit to OpenAI’s liability.

Similarly to Google’s indemnification, Open AI’s indemnification says that users will indemnify OpenAI against third-party claims arising out of or relating to the user’s use of Open AI’s platform and content or any violation of the terms. But then separate terms apply for user’s use of APIs, ChatGPT Enterprise, ChatGPT Team, and other services for businesses and developers, and those terms have mutual indemnification.

Platform: Adobe

Adobe provides users with AI tools primarily through its Creative Cloud suite, with specific licenses to use various generative AI features (like Adobe Firefly). Users must adhere to Adobe’s licensing terms and may not use AI-generated content in ways that infringe upon third-party rights. Adobe collects user data to improve its services, but data privacy and retention are guided by Adobe’s comprehensive privacy policies. Users are encouraged to review these terms to understand how their content may be used. Adobe’s AI tools cannot be used for generating harmful or illegal content. Violations may result in account termination and loss of access to services.

It’s important to note that other sections in the Adobe terms, including indemnification, apply based on your level of subscription. For users with a specific Creative Cloud plan, Adobe agrees to defend any third-party claims alleging that the user’s use of Firefly-generated output directly infringes on third-party copyrights, trademarks, publicity rights, or privacy rights. Otherwise, Adobe’s general terms of use apply, which state that users will indemnify Adobe from any claims arising from AI-generated content, use of the platform, of violation of terms. There is a limitation to Adobe’s liability as well.

Platform: CoPilot (GitHub and Microsoft)

CoPilot’s AI terms say that by using / uploading content, users give GitHub the right to use (and the right to sublicense to suppliers of services to GitHub) content for its own operation of business to improve service quality and offer new features. GitHub collects data to refine the CoPilot tool, but users maintain ownership of their code. Like other platforms, CoPilot prohibits users from generating harmful, illegal, or malicious code using its tools.

Users agree to indemnify GitHub for their use of the platform, particularly in relation to the generated code violating the rights of others or being used inappropriately. There does not appear to be any indemnification by GitHub.

In General: Indemnification

Most AI platforms have clauses that require users to indemnify the platform for third-party claims arising from user behavior, such as illegal use of the platform, intellectual property infringement, or misuse of data.

In some cases, platforms may provide indemnification in specific instances, such as if the AI provider is found to infringe on third-party intellectual property or is responsible for a service failure.

Conclusion

All AI platform terms of use generally agree that:

– The platform is provided as is.

– The platform’s liability is limited.

– Users cannot use the platform for illegal or harmful activity.

– Users are responsible for the content they input into the platform and the use of the content that AI outputs.

While there are similarities in the overall structure of indemnification clauses across platforms, the nuances of these provisions depend heavily on the specific nature of the AI tool, the level or type of subscription you have, and the potential legal consequences users may face. By reviewing these provisions carefully, users can ensure they are informed and better prepared to handle any legal challenges that may arise during their use of AI.

For now, we recommend that agencies approach AI cautiously. Be transparent with your clients about how and when you use AI, make sure you address AI in your contracts and are fully protected, and get your clients’ written consent. You want to make sure you clearly outline your AI policy with your clients and vendors so that your liability is limited in these engagements.

Questions about your agency’s AI use or how these platform terms apply to you? Reach out to us.