AI patent drafting: A blessing and a curse

From schools to workplaces, from solving complex mathematics problems to helping us remove grass stains from our jeans, artificial intelligence (AI) has found its place in our daily lives and we use it daily to navigate the challenges of living in the XXI century.

As AI-driven tools become more sophisticated, they are increasingly embedded in the workplace and the field of IP is not immune to this.

AI-based tools simplify the research, drafting and prosecution of patent applications, offering greater efficiency and speed, which is appealing to both patent firms and inventors.

However, these tools also introduce new challenges, particularly in terms of reliability as the information is now being processed by algorithms instead of a qualified patent attorney or paralegal.

Patent attorneys possess specialised knowledge of IP laws and are trained to apply it strategically and in line with the commercial goals of the applicant. While patent attorneys and paralegals use AI to assist them in daily tasks and helping in analysing documents, they still have a qualified look over the information produced by AI tools.

Inventors, by contrast, often approach AI as a source of specialised information, equating it to IP knowledge and understanding of the regulations surrounding patents.

This often introduces entropy in the drafting process, delaying feedback and agreement on finalised versions of the application.

This article discusses how AI influences the roles of patent attorneys and inventors, highlighting both its benefits and its risks in patent drafting.

AI and patent attorneys

Having the responsibility of preparing, filing, and prosecuting patents, patent attorneys play a critical role in the filing of new patent applications. Traditionally, this process is labour-intensive, requiring an understanding of both legal and technical aspects.

AI assists patent attorneys by automating routine tasks, enabling them to focus on more nuanced aspects of patent law.

AI-powered tools can conduct prior art searches more efficiently than humans, analysing vast databases swiftly to identify relevant existing patents or publications. This speeds up the preliminary assessment phase, providing attorneys with a solid foundation to build upon.

Additionally, AI tools can also summarise documents and extract critical points from extensive readings, saving time and making the process more efficient.

Finally, AI improves the drafting process by offering suggestions for definitions and the description, ensuring support for claims. AI also assists in organising the information in a patent draft, ensuring that all the information needed is present.

However, AI is not a substitute for drafting claims and in particular, it cannot extract the inventive concept from an invention disclosure. In fact, providing the claims and the technical effect of the essential features of the invention are the key requisites to start engaging with an AI tool, and these should be human made.

Furthermore, AI-based tools are not able to understand the impact of using one particular word instead of a seemingly synonymous word and its strategic implications for a better scope of protection.

In the end, and regardless of how AI is used, it does not dispense a qualified reading and checking of references to provisions in law and case law.

AI and inventors

Inventors are the minds behind new inventions and innovations, often resulting in breakthroughs that shape industries. AI enables access to patent-related resources and insights previously inaccessible without specialised knowledge or legal expertise which shortens the distance between the inventor and the patent attorney.

AI-driven platforms allow inventors to conduct initial prior art searches independently, giving them a preliminary understanding of existing patents and public domain technologies in their field. This early insight enables inventors to refine and differentiate their inventions before engaging with a patent attorney.

Moreover, AI helps inventors in drafting preliminary patent applications by providing templates and language suggestions, aligning their ideas with formal patent requirements. These tools simplify the initial stages of patent drafting, making it less confusing for those without legal backgrounds.

For inventors keen on optimising their patents’ potential commercial value, AI can analyse market trends and predict the commercial viability of their innovations. This data-driven approach assists inventors in making informed decisions about pursuing patents and potential market strategies.

However, AI-generated templates and suggestions can be particularly misleading. Although they may appear polished and authoritative, they often lack the nuanced legal reasoning required to ensure enforceability.

Inventors who rely too heavily on these tools may inadvertently omit critical details, misunderstand claim scope, or introduce inconsistencies that weaken their applications. This can create entropy in the patent drafting process, and especially in the middle of the drafting process, where the patent attorney provides versions of the patent draft to the inventor; relying on AI suggestions affects the efficiency of the process and may result in greater legal uncertainty.

When inventors subject patent drafts to AI tools instead of applying a critical reading to the draft, it often generates drafts that are unrelated to the original invention, leaving out critical elements of the claims. The dialogue between the inventor and the patent attorney no longer happens in a straight line.

Instead, it is being intermediated by a tool that is not critically evaluating the enforceability of the patent application, combining words and concepts in an unqualified manner.

In the end, as much as an AI algorithm cannot be an inventor, AI tools are not a replacement for an inventor critically reading the claims and assessing whether its creation is fairly reflected therein.

The integration of AI into patent drafting is undeniably transformative, increasing efficiency and reducing the workload of patent professionals. At the same time, it enables access to patent knowledge which, in the past, was difficult to find and understand.

There may come a time in the near future where an AI assistant can replace the patent profession but for now, without the guidance of a patent attorney, the misuse or overreliance on AI can put at risk obtaining meaningful protection for inventions.

This is a co-published article, originally published on WIPR.