The U.S. Patent and Trademark Office (USPTO) has implemented a ban on the use of generative artificial intelligence (GenAI) tools among its staff due to concerns about technology security and risks such as bias, unpredictability, and malicious behavior. This information was revealed in an internal memo from April 2023, which stated that while the USPTO is committed to innovation, it is still in the process of determining how to responsibly incorporate these tools into its operations.
USPTO employees are allowed to engage with generative AI models, but only within a controlled internal testing environment. Paul Fucito, the USPTO’s press secretary, explained that this AI Lab is being utilized to understand the capabilities and limitations of generative AI while developing AI-powered solutions for critical business needs. However, employees remain prohibited from using external AI programs like ChatGPT or Claude for their work and cannot use any output produced by these tools, such as generated images or videos. They can, however, use specific approved AI programs found within the agency’s own public database for patent-related searches.
The USPTO also recently entered into a $75 million contract with Accenture Federal Services to enhance its patent database with advanced AI search capabilities. Tasked with the protection of inventors and the awarding of patents, the USPTO advises on U.S. intellectual property policy, which includes operating within the guidelines of technological advancements.
During discussions at a Google-sponsored event, Jamie Holcombe, the chief information officer of the USPTO, acknowledged the challenges posed by government bureaucracy in the adoption of new technologies. He expressed the need for overcoming systemic hurdles such as lengthy budgeting and procurement processes that inhibit rapid integration of AI tools.
The USPTO’s ban on generative AI is not unique. The National Archives and Records Administration has also restricted the use of tools like ChatGPT on government-issued devices, although it simultaneously holds presentations encouraging staff to consider AI technologies as collaborative tools. Meanwhile, NASA has adopted a more cautious approach, banning chatbots for sensitive data while allowing experimentation with AI for code writing and research summarization, along with collaborations on satellite data accessibility.
For more details, visit the USPTO’s official site.
