Patently Strategic Podcast: Patents and AI

For many years, standard knowledge had most of us believing that automation and the inevitable rise of the machines would upend blue-collar industries first. However then AI had one thing to say about all of that. From passing medical licensing exams to writing code to even acing the Uniform Bar Examination, AI has change into society’s newest (and maybe most succesful!) change agent within the skilled office. At an astonishing tempo, it’s erasing all assumptions as to what industries shall be most impacted. With the latest developments centering across the understanding and technology of textual content and pictures – the core ingredients to patents – it’s turning into abundantly clear that patent practitioners and inventors are removed from proof against AI’s results and attain. However is all of it only a ticking time bomb of despair for the skilled IP trade? Or might there be some hope for fixing the problems that plague the system and even perhaps an innovation renaissance born from the quantum leaps being taken by this new wave of know-how? It’s time we had a speak about this.

Episode Overview: Present Instruments, Future Options

On this month’s episode, Dr. David Jackrel and Dr. Ashley Sloat lead a two-part dialogue with our all-star panel that begins with a deep dive into the current state of AI patent instruments after which strikes on to an exploration of how these instruments might finally present options for most of the issues plaguing the trade. The episode explores:

• ChatGPT-4.0 vs. professionals on core competencies
• Why AI is evolving so quickly
• AI issues and hallucinations
• AI and public disclosure threat
• AI implications for inventorship
• The present state of AI-assisted patent looking, proofreading, drafting (rule and LLM-based), and prosecution instruments
• AI’s potential future position within the patent system for fixing issues with the PTAB, search high quality, and the unsustainable bar

AI is Consuming the World

Our information feeds are overrun with tales of how AI is profoundly reshaping the world at a breakneck tempo and reaching outcomes that have been solely not too long ago the mere desires (or maybe threats relying on who you ask) of science fiction. Consultants are asserting that AI’s affect shall be bigger than the Industrial Revolution and that we may very well be lower than 5 years from the singularity – the second when AI can reason better than us and is not underneath human management.

However the place are we now? These massive language mannequin neural networks like ChatGPT, which predict the subsequent most certainly phrase based mostly on likelihood, are reaching outstanding leads to a number of the most complicated human contexts:

• Core human competencies with duties like speech recognition, handwriting recognition, picture recognition, studying comprehension, and language understanding have been, if not computationally inconceivable, actually properly under the benchmarks of human efficiency solely as not too long ago because the onset of the pandemic. AI is now surpassing human performance in most of the core competencies and several other of the pattern traces remain vertical.
• Open AI’s ChatGPT passed all three parts of the US medical licensing exam – and did so inside a snug vary.
• OpenAI has been used to create 3D video games by prompters with no data of sport improvement and a functioning website from a messy notepad sketch. GPT-4 can now write code in all main programming languages.
• Not solely did GPT-4 pass the bar exam, however it did so scoring within the ninetieth percentile, passing all sections of the UBE with an accuracy price of 74.5% – 9.5% increased than common human aspiring attorneys, and exceeding Arizona’s minimal passing rating, which is the very best threshold among the many 36 states and jurisdictions utilizing the bar examination.

The tempo at which that is taking place and the trajectory of the place it’s going are staggering – taking place much more shortly than our personal wetware and dated working programs can presumably comprehend.

Not Excellent…But

However for all of the hype, it’s nonetheless removed from excellent. Past extremely documented points with desires to hack computers and break up marriages, AI additionally presently suffers from a phenomenon referred to as hallucination. The time period is commonly used to explain situations the place an AI system produces content material that appears reasonable however is completely fabricated. For instance, producing textual content that appears coherent and contextually applicable however is completely fictional. This grew to become a real-world drawback for a few New York attorneys this previous June, who submitted a authorized transient that included six fictitious case citations generated by ChatGPT.

So, whereas there are nonetheless some glitches within the matrix, it’s necessary to not get complacent and overlook the capabilities when contemplating the tempo of the evolution. As talked about, ChatGPT-4 handed the bar within the ninetieth percentile. Its predecessor model, 3.5, solely scored within the fifth percentile. That’s one 12 months and a half tick of a model to go from failing to prime of the category. The harder consideration turns into not if however when to embrace these instruments, particularly given the complicated, nuanced, high-stakes nature of patents.

AI Patent Instruments State of the Union

There are numerous AI-assisted software program instruments already out there for practitioners. Half one of many panel dialogue offers a 30,000 ft view of the state of AI instruments for IP and provides a glimpse into how these instruments can, will, and in some circumstances already are, getting used. The group discusses private experiences with and insights into optimizing using AI-assisted patent looking, proofreading, drafting (rule and LLM-based), and prosecution instruments.

Can AI Save the Patent System?

We’ve spent a good quantity of airtime discussing the existential threats going through the patent system and the dire want for reform. Proposed legislative human options are likely to get complicated, messy, generally contentious, and at all times slowed down within the political equipment. However innovation doesn’t relaxation. Void of political options, what if we might leverage AI-based options to mitigate towards most of the circumstances that necessitate reform?

Within the second a part of the panel dialogue, our group explores how AI might increase the patent system and people whose providers help it, how search high quality may very well be improved, and the way a time-traveling PHOSITA armed with the time-stamped corpus of human data might remedy issues with invalidations round novelty, non-obviousness, and enablement.

Dialogue Panel

Ashley and David are joined in the present day by our at all times distinctive group of specialists together with:

Mossoff Minute: AI and Inventorship

On this month’s Mossoff Minute, Professor Adam Mossoff discusses the patentability of AI-generated works and innovations. We’re additionally publishing excerpts as short-form videos on Instagram Reels, YouTube Shorts, and TikTok.

Associated Listening and Studying

To additional discover the matters mentioned, see the next past episodes and sources: