Patent Strategy Podcast: Patents and AI

For decades, conventional wisdom has led most of us to believe that automation and the inevitable rise of machines will first upend blue-collar industries. But AI had something to say about all of this.From passing the medical license exam write the code AI has become society’s newest (and perhaps most capable) agent of change in the professional workplace, all the way to passing the Uniform Bar Examination. At an alarming pace, all assumptions about which industries will be most affected are being dispelled. Recent advances around the understanding and generation of text and images, the core elements of patents, are making it abundantly clear that patent practitioners and inventors are not immune to the influence and influence of AI. . But is it all just a ticking time bomb of despair for the professional IP industry? Or is there hope for solutions to the problems plaguing the system, and even innovation that will emerge from the quantum leaps of this new wave of technology? Is there even a Renaissance? It's time to talk about this.

Episode Summary: Current Tools, Future Solutions

In this month's episode, Dr. David Jackrel and Dr. Ashley Sloat lead a two-part discussion with an all-star panel. This discussion begins with a deep dive into the current state of AI patent tools, then moves on to explore how AI patent tools work. The tool could ultimately provide solutions to many of the problems plaguing the industry. In this episode we discuss:

Professional Comparison of ChatGPT-4.0 and Core Competencies
Why is AI evolving so rapidly?
AI problems and illusions
AI and public risk
How AI will impact inventors
Current state of AI-assisted patent search, proofreading, drafting (rules and LLM-based), and examination tools
Potential future role of AI in patent systems to solve problems of PTAB, search quality, and unsustainable standards

AI engulfs the world

Our newsfeed is filled with articles about how AI is profoundly reshaping the world at breakneck speed, achieving results that until recently were only science fiction dreams (or threats, depending on who you ask). It's overflowing. Experts say the impact of AI will be greater than the industrial revolution, and that the singularity (the moment when AI will be able to reason smarter than humans and will no longer be under human control) could be less than five years away. claims.

But where are we now? Neural networks in large-scale language models like ChatGPT predict the next most likely word based on probability and have achieved remarkable results in some of the most complex human contexts.

core human abilities Tasks such as speech recognition, handwriting recognition, image recognition, reading comprehension, and language understanding, if not computationally impossible, have fallen significantly below human performance standards for the first time as recently as the onset of the pandemic. I'm sure it was. AI currently outperforms humans in many core competencies, and some trend lines are still vertical.
medicine. Open AI's ChatGPT passed all three parts of the U.S. medical licensing exam, and did so within a comfortable range.
software. OpenAI has been used to create 3D games by teleprompters with no knowledge or expertise in game development. A functioning website From a sketch on a dirty notepad. With GPT-4, you can now write code in all major programming languages.
Law. GPT-4 not only passed the bar exam, but also scored in the 90th percentile and passed all sections of the UBE with a 74.5% correct answer rate. This put him 9.5% higher than the average for human lawyer applicants and higher than Arizona's minimum passing score. , which is the highest threshold among the 36 states and jurisdictions that use the bar exam.

The pace at which this is happening and where it's going is amazing, and it's happening much faster than our wetware and outdated operating systems can possibly comprehend.

It's not perfect…yet.

But despite all this hype, it's still far from perfect. Aside from its many documented problems with hacking computers and ruining marriages, AI also currently suffers from a phenomenon known as hallucinations. This term is often used to describe when an AI system produces content that looks realistic but is completely fabricated. For example, it produces text that appears coherent and contextually appropriate, but is completely fictitious. This became a real issue for several New York lawyers in June of this year, who filed a legal brief containing his six hypothetical case citations generated by ChatGPT.

So while the Matrix still has some glitches, it's important not to get complacent and overlook features when considering the pace of evolution. As mentioned earlier, ChatGPT-4 exceeded the 90th percentile criterion. The previous version, 3.5, only gave you a score in the 5th percentile. This equates to a year and a half of time from failure to reaching the top of the class.Eliminates the need for more difficult considerations if but when These tools should be leveraged, especially given the complex, nuanced, high-stakes nature of patents.

AI Patent Tools State of the Union

There are many AI-assisted software tools already available to practitioners. Part 1 of the panel discussion will provide a 30,000-foot view of the current state of AI tools for IP, and discuss how these tools can, will, and may be used. You can get a glimpse of what is already in use. The group will discuss personal experiences and insights for optimizing the use of AI-assisted patent search, proofreading, drafting (rules and LLM-based), and examination tools.

Can AI save the patent system?

We have spent considerable air time talking about the existential threats facing the patent system and the urgent need for reform. Proposed legislative human solutions are complex, cumbersome, sometimes controversial, and always tend to get bogged down in political institutions. But innovation never rests. In the absence of political solutions, what if AI-based solutions could be leveraged to alleviate many of the conditions that require reform?

Illustration: Declan Rede

In the second part of the panel discussion, our group discussed how AI can enhance the patent system and the services that support it, how it can improve search quality, and how AI can improve the quality of time-stamped human Consider how a time-traveling PHOSITA with a corpus of knowledge would work. It may resolve invalidation issues related to novelty, nonobviousness, and validation.

discussion panel

Ashley and David today join an all-time great group of professionals including:

Mosov Proceedings: AI and Inventiveness

In this month's Mosoff Proceedings, Professor Adam Mosoff discusses the patentability of AI-generated works and inventions. He also publishes excerpts as short videos on his Instagram Reels, YouTube Shorts, and TikTok.