Adobe, Microsoft Make Separate Generative AI Pledges

Adobe and Microsoft have separately announced that they will help protect users of their generative AI products from legal challenges, adding some much-needed responsibility to our emerging AI era.

Adobe’s approach is much more substantive and simpler. It is offering customers “full indemnification for the content created through” its generative AI capabilities in Adobe Firefly, which is now rolling out to business users, Adobe vice president Claude Alexandre said today at the firm’s Adobe Summit today. In other words, Adobe will legally compensate any customers that are sued for copyright infringement for images created with Firefly.

And on that note, Adobe is rolling out Firefly-based AI capabilities today in Photoshop (beta), so customers can start using natural language prompts to generate images. One can also access this capability directly in the Firefly beta. You can learn more about these capabilities on the Adobe Blog.

Microsoft isn’t offering indemnification. Instead, it is instituting an AI Assurance Program to help ensure that the AI applications its customers deploy on its platforms meet the legal and regulatory requirements for responsible AI.

“Microsoft has been on a responsible AI journey since 2017, harnessing the skills of nearly 350 engineers, lawyers and policy experts dedicated to implementing a robust governance process that guides the design, development and deployment of AI in safe, secure, and transparent ways,” Microsoft corporate vice president and deputy general counsel Antony Cook says of his firm’s efforts. “We are creating an AI Assurance Program to help you ensure that the AI applications you deploy on our platforms meet the legal and regulatory requirements for responsible AI.”

Microsoft has adapted an approach it took previously with the financial industry to create a “know your customer” (KYC) strategy that “creates certain obligations to know one’s cloud, one’s customers, and one’s content.” The firm says it will work with its customers to apply KY3C as part of its AI Assurance Program, share its AI Risk Management Framework implementation, instigate customer councils to address feedback on AI usage, and ” play an active role in engaging with governments to promote effective and interoperable AI regulation.”

As Microsoft notes, these commitments are only a start, and much more is needed. But I doubt that the software giant will ever be able to offer legal indemnification given the public sources of data it uses to generate its AI models.