Internet & Social Media Law Blog
For users of generative AI programs, a growing concern has been with potential liability resulting from infringement claims by copyright owners whose materials were used to train the AI. At its annual DevDay conference in early November, OpenAI became the latest major company to address this by offering to indemnify certain users of its ChatGPT chatbot.
Generative AI has seen a meteoric rise in recent years, with applications spanning from content creation to software development. However, this rise has been accompanied by growing intellectual property (IP) concerns. A core issue is determining who bears responsibility for AI-generated content: the user or the AI entity itself?
We discussed this issue on my panel at this year’s annual meeting of the Intellectual Property Owner’s Association (IPO). There, our focus was mainly on image-generating AI. We recommended using programs offered by companies that would stand behind their programs and agree to indemnify users against liability for infringement claims.
For example, Adobe and Shutterstock have introduced their own forms of legal protection for users of their image-generating AI tools. Adobe has begun offering IP indemnity for commercial users of Firefly, its generative AI-powered image creator application. Shutterstock has also introduced an indemnity policy for its enterprise-level customers, covering the use of AI-generated images created on its platform.
Other major players in AI have also rolled out similar indemnity policies. Microsoft-owned GitHub was among the first to offer indemnity for its Copilot software-development AI, provided that users utilized Copilot’s “duplication detection” filter. IBM and Google have also recently announced policies to indemnify users of their respective AI platforms. These protections, while subject to various terms and conditions, represent a significant step in the direction of user-focused liability shields for AI-generated content.
With its announcement on November 6, OpenAI became the latest major AI company to offer indemnity for certain users of its ChatGPT platform. A corresponding blog post on OpenAI’s website claims that the company will “step in and defend our customers, and pay the costs incurred, if [they] face legal claims around copyright infringement.”
Yet, OpenAI’s new policy—dubbed the “Copyright Shield”—is not without limits. It only applies to enterprise-level ChatGPT customers and users of OpenAI’s developer platform. Further limitations are laid out in the terms of use. That said, it appears to be among the more comprehensive indemnity policies from a generative AI company. It follows other changes OpenAI has made to add more “built-in copyright safeguards” to the platform, as the company and its peers have faced a rash of high-profile lawsuits alleging that their programs rely on and facilitate widespread copyright infringement.
To be clear, the industry standard for generative AI companies is still to disclaim liability for the output of their AI platforms. Even these companies that have started offering indemnity place significant limitations on who qualifies for protection—typically, only the highest-paying subscription tiers. Nonetheless, these moves by prominent AI companies mark a significant shift. It may result in smaller companies feeling the pressure to adopt similar protections for their users in order to remain competitive.
As the market continues to evolve, the legal landscape will undoubtedly become more complex. Users and developers must navigate these waters with a strategic mindset, ensuring they are both protected and compliant. For now, with the indemnity being offered by OpenAI, Google, Adobe and other major players, the generative AI sector inches closer to a future where legal clarity and user confidence may go hand in hand.