Standards

• Auto-generated Legal Documents 

  • State of the art machine learning systems are capable of generating news articles and documents that are often indistinguishable from those created by human authors. Systems that generated legal documents, in whole or in part, could potentially save time, reduce errors, and improve quality. Nevertheless designing systems capable of useful legal drafting requires that we first specify several concrete objectives for those systems. For example, what exactly do we mean by quality? Can we employ any objective metrics of legal drafting quality? In addition, automated legal drafting raises many other design decisions, such as which drafting tasks to delegate to software and which should remain within the purview of human drafters, and how a desire to standardize legal documents shapes automatically drafted texts.  

     
• Application Programming Interface (API) to Patent Analytics Software
  • Patent analysis software, patent search engines, and other software tools are being increasingly used by law firms, corporate patent departments, and others in the patent community. Firms and law departments are also creating their own tools to focus on the data most relevant to their needs. As the creation of custom patent software becomes more common, the patent community would benefit from an API to existing patent software provided by large and small software providers. This would allow the developers of custom software to benefit from standardization, more rapid development, and other well-known benefits of APIs. 
• Prior Art Search Metrics and Benchmarks

  • Community standards have a history of boosting innovation and the pace of development in an industry. For example, the HTTP, HTML, and URL standards were instrumental in the rapid expansion and adoption of the Internet and the World Wide Web. AI patent search tools continue to increase in power and capabilities. However, development on AI patent search technologies is scattered among numerous companies that duplicate effort. Unlike in many industries, AI patent tools lacks initiatives to coordinate or distribute work on the largest, most valuable objectives. Consequently, potentially transformative patent search functionality is neglected because it is too challenging for any single company to create. Open-source metrics and benchmark data sets for these metrics would benefit users of patent search tools.  

 If you're interested in working on any of these projects, please contact us.