“Striking the Perfect Balance: A Guide to Ensuring Trust, Transparency, and Truthfulness in Government”

Why the US Patent and Trademark Office has Decided Against the Use of Large Language Models in Government Activities

With the rapid development of technology, government agencies are faced with the challenge of adopting and using emerging technologies while balancing the benefits with the associated risks. One such emerging technology that is currently taking the world by storm is Large Language Models (LLMs) with their endless potential use cases. However, as with any emerging technology, there are potential risks and downsides to consider before moving forward.

The United States Patent and Trademark Office (USPTO) is known for its role in safeguarding intellectual property rights, and recently made a significant decision by choosing not to use generative AI and Large Language Models (LLMs), including ChatGPT. At the April 2023 GovFuture Forum event held at George Mason University in Washington, DC, Scott Belevieu, Branch Chief of Advanced Analytics & Acting Director of Data Architecture in the Office of the Chief Technology Officer (OCTO) at the USPTO shared his agency’s perspective on the use of AI and LLMs and their current posture with regards to their use within government activities.

The Growing Use of AI

Artificial Intelligence (AI) is now widespread in many aspects of daily life, with most people unknowingly using AI for everyday tasks such as recommending products or TV shows, or even drafting emails and writing articles. AI has the potential to provide tremendous societal and economic benefits, including fostering a new wave of innovation and creativity. According to Belevieu, AI now appears in 18% of all new utility patent applications and in over 50% of all the applications that the USPTO examines.

However, AI-generated content, including LLMs, also introduces numerous challenges and risks, particularly in terms of transparency, truthfulness, and trust. The lack of clarity surrounding AI algorithms and the potential biases embedded within them causes concerns about transparency. Truthfulness and trust are also critical issues, particularly within the patent examination process, where the potential for misinformation poses a significant challenge. For these reasons, the USPTO is highly cautious about using the technology.

Navigating Trust, Transparency, and Truthfulness

The USPTO, like other government agencies, has a responsibility to ensure the highest level of integrity, fairness, and accountability, especially when dealing with intellectual property matters. Given the potential limitations and biases inherent in AI systems, maintaining public trust and transparency is paramount.

Belevieu categorized the risks associated with AI and emerging technologies into three buckets: trust, transparency, and truthfulness. In the public sector, it’s essential for government officials to maintain the trust of the public. This trust is threatened if the models generate content that is biased, malicious, or not based on reputable sources. The second risk is transparency. The IP system fundamentally relies on trust and transparency. To ensure this, there needs to be openness to know that decisions are based on facts that led to that particular decision. Finally, truthfulness is the third risk. AI-generated content, including LLMs, can sound truthful, but the information or the basis of a recommendation may not have a solid foundation.

USPTO’s Decision

The USPTO’s decision not to adopt generative AI and LLMs is rooted in safeguarding trust, transparency, and truthfulness. By avoiding these technologies, the USPTO demonstrates its commitment to ensuring fairness and avoiding potential biases or uncertainties associated with AI systems. The decision aims to uphold the agency’s dedication to maintaining public confidence in the patent examination process.

Conclusion

The use of AI and LLMs pose a significant challenge for government agencies, and there is a constant risk of potential biases and inaccuracies. The US Patent and Trademark Office has decided to avoid these technologies, demonstrating a commitment to trust, transparency, and truthfulness in their intellectual property procedures. As AI technology continues to develop, it is crucial to prioritize openness and transparency to maintain public trust in government activities.