The advent of the internet has been one of the most significant technological advances: artificial intelligence AI, a transformative technology with enormous economic and societal benefits. Integrating AI technology into next-generation tools presents an exciting opportunity for improving trademark and patent examination quality and efficiency. The USPTO has been utilizing artificial intelligence (AI) in various aspects of its operations to improve the efficiency, accuracy, and speed of its processes.
AI-assisted Error Detection
Patent Examination: AI-powered systems can assist examiners in reviewing patent applications by quickly searching through large databases of prior art and highlighting relevant information. This can speed up the examination process and improve the accuracy of patent grants.
Image Processing: AI can be used to extract information from images in patent applications, such as drawings and diagrams, to help examiners better understand the invention described.
Patent Analytics: The USPTO uses AI to analyze patent data and identify trends, helping the office to make data-driven decisions and improve its operations.
Patent Classification: AI can be used to automatically classify patent applications into specific categories, making it easier for examiners to review them and ensuring that they are assigned to the correct examination team.
Overall, the USPTO’s use of AI is helping to improve the speed and accuracy of its processes, making it easier for inventors to get their patents granted and enabling the office to better serve its stakeholders.
AI-assisted Patent Examination
A thorough prior art search is an essential part of patent examination and the USPTO’s mission to grant reliable patent rights. It is becoming increasingly difficult to find the most relevant prior artwork due to technological advancements and the rapid growth of prior art. The USPTO has created an AI-based prototype search tool that can help users identify relevant documents and provide suggestions for other areas to search. The system provides world-class patent AI models and is designed to learn from its USPTO examiners, who are some of the most respected patent searchers in the world. To provide additional enhancements, the system automatically collects feedback data from its examiners. To provide transparency, the USPTO is developing features that will allow examiners to interpret the AI models’ results. In March 2020, a beta version of the new AI tool was made available to select examiners. Evaluations have shown promising results so the USPTO is taking steps to integrate AI into the next generation of search tools for examiners.
a.Define your invention.
Clearly state the main aspects and features. This will narrow down your search and help you identify relevant prior art.
b.List keywords and phrases that are relevant to your invention.
This can include technical terms or product names. It could also be specific components.
c. Patent databases can be a great resource when searching for prior art.
Use platforms such as the United States Patent and Trademark Office, the European Patent Office or WIPO Patentscope. Enter keywords to explore patent documents.
Classification and Categorization
An auto-classification tool, which leverages machine learning to classify patent documents under the Cooperative Patent Classification system (CPC), was also developed. The system can suggest CPC symbol suggestions and allows for the identification of claimed subject matter to refine the suggested CPC symbol suggestions. This is similar to its AI-search system.
Indicators that link suggested CPC symbols with specific parts of the document provide insight into the reasoning of an auto-classification system. The system includes enhanced feedback mechanisms that can be integrated with existing classification processes in order to aid in training the AI. The USPTO used system performance analysis to implement auto-classification in December 2020. This allows the USPTO to automatically identify CPC subject matter for internal operations. The agency has seen a reduction in its procurement costs for CPC data acquisition. The USPTO is also developing new capabilities to help support a wider range of patent classification requirements at USPTO.
AI Usage at the PTO
The USPTO recently conducted market research on AI capabilities for image comparison, and to verify the acceptance of services and goods being identified against entries in the Trademarks ID Manual. USPTO developed AI prototypes that could be used to compare trademark images and suggest the right assignment of mark design codes. This would also help to assess the acceptability of identifications of goods or services.
These prototypes were tested through a common interface with around 10 stakeholders in November 2020. A larger beta is possible later in the year. The USPTO also tested false specimen detection solutions using a software program. This software program was integrated into agency efforts to identify digitally altered specimens of use and mock-up web pages on December 1, 2020. A prototype of an AI-based chatbot that answers frequently asked questions via USPTO’s website may be available for beta testing in the latter part of this year.
It’s likely that the US Patent and Trademark Office (USPTO) will continue to explore and adopt new artificial intelligence (AI) technologies in the future to improve the efficiency, accuracy, and speed of its processes. Here are a few potential areas where the USPTO could use AI in the future:
AI-powered translation systems could be used to automatically translate patent applications and related documents into multiple languages, making it easier for examiners to review and understand inventions from around the world.
AI could be used to analyze patent data and make predictions about the likelihood of a patent being granted or the likelihood of a patent dispute, helping the USPTO to allocate resources more effectively and make data-driven decisions.
AI-powered systems could be used to detect fraud and inconsistencies in trademark applications, helping the USPTO to ensure the integrity of the trademark system and prevent fraud.
AI-powered virtual assistants could be used to answer common questions from inventors and the public, freeing up USPTO staff to focus on more complex tasks.
Overall, the future use of AI by the USPTO will depend on the development and availability of new AI technologies, as well as the office’s ongoing efforts to improve its processes and better serve its stakeholders.
How AI is Used to Reject Patent Applications
Artificial intelligence (AI) can be used to assist in the rejection of patent applications. AI systems can be trained to analyze patent applications and identify inconsistencies, inaccuracies, or prior art that may render an invention unpatentable. This can help speed up the patent examination process and improve the accuracy of patent grants by identifying problematic applications more quickly.
A rejection under 35 U.S.C. § 101 of the US Patent Act occurs when a patent application is rejected because the claimed invention is deemed to be not eligible for patent protection. This can occur when the claimed invention is deemed to be directed to a law of nature, a natural phenomenon, an abstract idea, or a purely conventional or obvious method.
Artificial intelligence (AI) can be used to assist in the rejection of patent applications under 35 U.S.C. § 101 by analyzing patent applications and identifying potential eligibility issues. AI systems can be trained to analyze the content of patent applications and identify potential abstract ideas or natural phenomena, as well as determine whether a claimed invention is directed to a conventional or obvious method.
AI) can be used to assist in applying the Alice factors in a rejection under 35 U.S.C. § 101 of the US Patent Act. The Alice framework, named after the Supreme Court case Alice Corp. v. CLS Bank International, is used by the US Patent and Trademark Office (USPTO) and the courts to determine whether a claimed invention is eligible for patent protection under 35 U.S.C. § 101.
The Alice framework involves a two-step process for evaluating patent eligibility: (1) determining whether the claimed invention is directed to an abstract idea, and (2) determining whether the claimed invention contains an inventive concept sufficient to transform the abstract idea into a patent-eligible invention.
AI can be trained to analyze patent applications and identify potential abstract ideas, as well as determine whether the claimed invention contains an inventive concept that transforms the abstract idea into a patent-eligible invention. This can help speed up the patent examination process by identifying problematic applications more quickly and improve the accuracy of patent grants by reducing the number of applications that are granted in error.
It’s important to note that the final decision to reject a patent application under 35 U.S.C. § 101 still rests with a human patent examiner. AI systems can only provide assistance and support to examiners by highlighting relevant information and potentially problematic areas, but the ultimate decision must be made by a person who is knowledgeable about patent law and who can carefully consider all relevant factors.
The USPTO has been incorporating artificial intelligence (AI) into various aspects of its examination process. AI systems are already in use to assist patent examiners with filtering out unpatentable applications, similar to the way AI systems help the pharmaceutical industry screen drug candidates.
According to the USPTO, AI systems could be used to examine and analyze patent application documents, allowing USPTO examiners to spend less time examining applications and more time responding to those that qualify for a patent. This process can be helpful for both applicants and examiners by helping to lighten the workload on examiners as well as increase the quality of the patent applications that are allowed.
Aside from the question of authorship, other issues to consider in AI cases include whether or not the inventive subject matter around which an AI invention was conceived is exclusively attributable to an AI system, or if it was primarily created by an individual. The USPTO has issued guidance indicating that certain inventions involving machine learning applications are patent eligible, but it is important to take the time to evaluate inventorship in these cases.
A rejection under 35 U.S.C. § 102 of the US Patent Act occurs when a patent application is rejected because the claimed invention is deemed to be anticipated by prior art. In other words, the claimed invention is not considered novel or non-obvious in light of prior disclosures or prior patents.
As a rule, 35 U.S.C. §102 requires that any claim of an invention be new or novel (and not obvious to a person having ordinary skill in the field). This requirement can be difficult to overcome, especially for a complex or advanced technology like artificial intelligence. However, using AI to examine new patent applications can make the examination process more efficient. This may help patent applicants and the PTO ensure that new technologies receive valid patents. It can also increase the value of issued patents, which can lead to more technological investment and commercialization.
Artificial intelligence (AI) can be used to assist in the rejection of patent applications under 35 U.S.C. § 102 by analyzing patent applications and identifying relevant prior art that may anticipate the claimed invention. AI systems can be trained to search large databases of patents and other technical documents, making it easier and faster to find relevant prior art.
A rejection under 35 U.S.C. § 103 of the US Patent Act occurs when a patent application is rejected because the claimed invention is deemed to be unpatentable due to obviousness in light of prior art. In other words, the claimed invention is considered to be an obvious combination or variation of existing elements or inventions, and therefore is not considered to be novel or non-obvious.
Artificial intelligence (AI) can be used to assist in the rejection of patent applications under 35 U.S.C. § 103 by analyzing patent applications and identifying relevant prior art that may render the claimed invention obvious. AI systems can be trained to search large databases of patents and other technical documents, making it easier and faster to find relevant prior art.
However, it’s important to note that the final decision to reject a patent application under 35 U.S.C. § 103 still rests with a human patent examiner. AI systems can only provide assistance and support to examiners by highlighting relevant information and potentially problematic areas, but the ultimate decision must be made by a person who is knowledgeable about patent law and who can carefully consider all relevant factors.
One of the most frequent patent rejections issued by the USPTO is a Section 112(b) rejection. This rejection occurs when the patent examiner considers a claim element to be “indefinite.” A Section 112(b) rejection can often be remedied by amending the patent application, ensuring that all terms are well-drafted and supported by the specification.
AI is being used for detecting errors in patent specifications, including erroneous parts lists and numbering inconsistencies. Patent specifications typically include a detailed description of the invention, including drawings, diagrams, and other illustrations.
AI systems can be trained to analyze patent applications and identify inconsistencies or inadequacies in the written description or enablement of an invention, which could result in a Section 112 rejection. This can help speed up the patent examination process by identifying problematic applications more quickly and improve the accuracy of patent grants by reducing the number of applications that are granted in error.
It’s important to note that, as with any use of AI in the patent examination process, the final decision to reject a patent application under Section 112 still rests with a human patent examiner. AI systems can only provide assistance and support to examiners by highlighting relevant information and potentially problematic areas, but the ultimate decision must be made by a person who is knowledgeable about patent law and who can carefully consider all relevant factors.
The Future of AI at the USPTO
The future of AI at the United States Patent and Trademark Office (USPTO) holds immense potential for revolutionizing the patent and trademark examination processes. As technology continues to advance, we can anticipate several exciting developments.
One key area of progress lies in advanced prior art search capabilities. AI tools will become increasingly sophisticated in their ability to analyze vast amounts of data and identify relevant prior art with greater accuracy and efficiency. This will enable examiners to make well-informed decisions, reducing the chances of granting patents for inventions lacking novelty.
Enhanced language processing will also play a significant role. Natural language processing algorithms will continue to evolve, empowering AI systems to comprehend complex legal and technical language used in patent and trademark documents. This will improve the extraction of key information, semantic analysis, and the delivery of more accurate and insightful results to examiners.
Furthermore, the future of AI at the USPTO may involve intelligent patent classification systems. AI algorithms will be trained to automate the categorization of patent applications, streamlining the application management process, ensuring consistency, and expediting the examination workflow.
Another exciting prospect is the integration of AI into decision-making processes. By leveraging vast amounts of historical data, AI models can provide decision support and predictive analytics to examiners. These tools will assist in assessing the likelihood of patentability, potential conflicts, and infringements, empowering examiners to make more informed decisions.
Moreover, the USPTO may explore collaborative tools and crowdsourcing mechanisms powered by AI. These platforms would facilitate knowledge sharing and collaboration among examiners, leveraging AI algorithms to identify experts, enable discussions, and tap into crowdsourced expertise for thorough examination and evaluation of complex inventions.
In addition, proactive detection of IP infringements will become a vital area of focus. AI systems can monitor diverse sources, including patents, trademarks, product descriptions, and online content, to identify potential conflicts and notify relevant stakeholders. This proactive approach will help safeguard intellectual property rights more effectively.
Lastly, AI can streamline the application process itself. AI-powered systems can provide real-time guidance, automated form filling, and error detection, simplifying and expediting the application process for patent and trademark applicants
PowerPatent’s Tools for Patent Applicants
PowerPatent is a company that provides a range of tools and services to help patent applicants in the preparation and filing of their patent applications. The goal of PowerPatent is to even the playing field for patent applicants by providing them with affordable access to high-quality patent preparation and prosecution services.
PowerPatent’s tools can help patent applicants in various ways, including:
- Providing guidance on the preparation and drafting of patent applications, including assistance with claim language, drawings, and other important components of the patent application.
- Automating certain tasks, such as parts list and numbering consistency checks, to help identify and prevent errors in the patent application.
- Providing assistance in the preparation of responses to office actions from the US Patent and Trademark Office (USPTO) and other patent offices.
- Providing online tools for tracking the status of pending patent applications and for managing communication with patent examiners.
By providing these tools and services, PowerPatent aims to make the patent application process more accessible and more affordable for inventors, entrepreneurs, and small businesses, helping to ensure that their innovations are properly protected. However, it’s important to note that while PowerPatent’s tools and services can be a valuable resource for patent applicants, they do not guarantee the grant of a patent, which is ultimately determined by the USPTO and other patent offices based on their own policies and regulations.
The USPTO has embraced the use of AI tools to support patent and trademark examination. These tools utilize machine learning, natural language processing, image recognition, and predictive analytics to assist examiners in various aspects of their work. By leveraging AI, the USPTO aims to enhance search capabilities, automate classification, process image-based trademarks, analyze language and semantics, and even predict application outcomes. These advancements help improve the efficiency, accuracy, and consistency of the examination process, ultimately benefiting inventors, businesses, and the public. The deployment of AI tools by the USPTO represents a significant step forward in harnessing technology to optimize intellectual property examination and management.