Patent Lodge

We are pleased to introduce our colleague and partner, Steve Helland. Steve has long provided guidance to clients in legal matters involving information technology, Internet, software, new media and advertising. Steve chairs Fredrikson’s Tech & Data legal group, which further advises on legal issues relating to data privacy, cybersecurity, technology and intellectual property issues.

In recent years, Steve has expanded his practice to focus on the exploding area of artificial intelligence, and to guide clients on related legal issues. We are happy to introduce Steve as a potential resource for you and your clients. Take it away, Steve!

            I’m writing this in February 2024—and that is important to capture, as each month we see changes in how various Generative AI tools, such as ChatGPT, Bard, Bing, Claude, and others perform, how early courts and government agencies provide decisions and guidance, and how individuals, business, and other institutions adapt. With that caveat, here are some items that should be front of mind for attorneys advising on these issues:

• Everybody’s using it. Consider adopting a policy. In my experience, use of Generative AI tools by employees across all industries and at all levels of seniority is extremely common. Friends and contacts quietly admit to all sorts of uses, often with little or knowledge and/or guidance from the employer. Consider adopting a Generative AI Use Policy that is a good fit for your business, industry, use cases and risk-tolerance.
• The “raw” output of Generative AI tools is not copyrightable or patentable. Both the US Copyright Office and the USPTO have issued guidance that the raw, unaltered output of a Generative AI tool is not copyrightable and may not give rise to a patentable invention. How much human input or modification is required remains an open question.
• All Generative AI tools are not the same. The performance and quality of output is not the same. The legal terms and conditions appliable to different Generative AI tools, and different versions of those tools (e.g., paid version vs. free) are not the same. The underlying “training data” used to train the tools is not the same.
• Be careful with your confidential information. Inserting company confidential information into a prompt window may jeopardize the confidentiality of that information.
• Not ready to fly solo. Given the continuing frequency of factual errors and other problems with inappropriate output from generative AI tools, humans need to continue to supervise and stay in the loop on all (or virtually all) use cases. For attorneys themselves, this is an ethical requirement.
• Stay tuned! We are just beginning this journey—with expected substantial technical, legal, business and cultural change ahead!