Generative AI is one of the latest and most revolutionary applications of AI technology, with the ability to generate content ranging from text to video using neural networks that are trained on predefined (and even user inputted) content. Generative AI is based on the concept of Generative Adversarial Networks (GANs), a concept developed by Ian Goodfellow and colleagues in 2014. The purpose of a GAN is to use a discriminative network to evaluate the outputs of a generative network, using the discriminative network to distinguish the generated data from the true data distribution that the model is trained on. As this process continues,the outputs of the generative network are trained by the responses of the discriminative network such that the outputs of the generative network are able to closely replicate the true dataset’s characteristics. With the power of GANs and generative AI, discriminative and generative functions of this new AI technology can assist the patent drafting process by analyzing prior art and patent databases to decrease the possibility of invalidity or infringement exponentially compared to a human-only drafting process.
[P1:identifying relevant areas to file patent claims and useful supplements for invention disclosures]
The first issue that non-AI assisted patenting systems and organizations face is keeping up to date with rapidly expanding patent filings. In business areas such as drug development and consumer retail, design patents play an instrumental role in protecting innovations. However, due to the IP-heavy strategy that is employed in pharmaceuticals and consumer goods, there is increased complexity in evaluating patent databases and prior art when crafting a company’s IP portfolio. With advanced webscraping and database integration, AI systems can more effectively track and identify areas to file patent applications and streamline the innovation process. The benefit of AI-powered systems is their ability to be trained on databases of existing data. By training software on patent databases that are continually updated, AI-powered IP management software solutions such as Patlytics are able to accurately guide IP specialists within companies regarding what additional patents or IP need to be created to protect a company’s product. Patlytics utilizes generative AI that is trained both on USPTO and WIPO database, alongside user inputs, enabling accurate and efficient patent generation.
[P2: improvements on infringement analysis, focused analyses of IP portfolios to suggest iterative improvements, hallucination risks to be mitigated with increasing usage]
Additionally, with this same integration, AI-powered IP management software also has the ability to detect infringement much faster than human teams and non-AI powered database search software alone. With analytical specificity provided by user-inputted patent and prior art data points, infringement reports can be generated with higher accuracy and increased efficiency. Additionally, AI systems can alert patent holders about potential infringement through continual database monitoring. With both database analysis and user querying of AI algorithms, hallucinations can be prevented by comparing AI output with user feedback and database information. While the risk of hallucination is not completely eliminated with this kind of software, the ability of users to review output and the feedback loop between users, the software, and databases increases the accuracy of the system as it is used. Additionally, by making data more digestible than human summary can (given the volume of data to be processed in the case of intellectual property), AI-powered systems incentivize users to utilize systems and thus increase their ability to generate more useful outputs.
[P3: conclusion- improved efficiency and ROI using AI patent drafting software, and risks to be mitigated with increasing usage + database training]
Intellectual Property is the cornerstone of business success. Without proper IP protection, businesses can lose protection of their key assets and face the prospect of having their innovation stolen by competitors. However, legal advising comes at a prohibitive cost for many firms due to the hours and manpower required to continuously monitor IP portfolios and overall IP landscapes for a given industry. AI-powered IP portfolio management software allows companies to increase the accuracy and scope of their intellectual property management by narrowing their portfolio and competitive focus to the tailored output of AI models trained on patent data and continuously updated IP databases. By streamlining the process of infringement, invalidity, and filing analyses, AI-powered IP management software increases the ROI for companies that utilize it, and also increases the impact that legal advisors and firms can have for their clients. Generative AI may be a developing technology, but in the hands of the right people and with the right data to train it, generative AI can make lawyers more efficient and businesses more successful.
