This week, the Department of Commerce announced that it was rescinding and would not enforce the Biden-era AI diffusion rule that was set to come into effect.  The AI diffusion rule aimed to regulate the global distribution of advanced AI chips and related technologies by implementing a tiered framework.  That framework would have allowed close allies of the United States to have the broadest access to advanced AI chips, given many countries more restricted access, and substantially limited access to countries of concern.  The Bureau of Industry & Security plans to issue a replacement rule in the future.  While the specific contours of that replacement are unknown, the Under Secretary of Commerce for Industry and Security Jeffery Kessler stated: “The Trump Administration will pursue a bold, inclusive strategy to American AI technology with trusted foreign countries around the world, while keeping the technology out of the hands of our adversaries."  It appears that the replacement will allow for broader global distribution of advanced AI technology.  

How will this impact IP?  

Perhaps the most profound impact will be on trade secrets.  AI advances level the playing field for developing technology that required significant time and effort to develop.  And AI coding capabilities are improving at a staggering pace.  Anthropic's CEO recently predicted in March that AI might be writing most code within 6-9 months.  Trade secrets that can be reverse engineered or independently created with use of AI will become less valuable as global development of AI increases.  

At the same time, unique data and secrets that cannot be independently created by throwing more compute and AI capabilities at a problem will become more valuable, so long as they are successfully protected as trade secrets.  Secret information such as clinical trial results in healthcare, geospatial data, customer behavioral data, or supply chain logistics data cannot be independently generated by AI.  That secret data in combination with improving AI capabilities could provide its holder with an even greater competitive edge.  With AI, the trade secret holder can draw new inferences from the data to offer new and better products and services.  And the trade secret holders can use the data to train AI models or license the data or trained models to others.  

An increase in global AI capabilities will also impact other aspects of IP.  It will enable new inventions, especially by independent inventors, small companies, and others who move quickly to adopt and leverage new AI capabilities.  Assistive use of generative AI will also accelerate the creation of copyrighted works.  

A key takeaway is to develop an IP strategy now with an eye toward where things are heading. Mark Cuban likely got it right last month when he predicted that “generative AI will be the greatest growth and productively engine, ever” and “intellectual property will be king as people try to determine what we value in the new world.”  But not all IP is equal - the key will be identifying and focusing on the IP that will be valuable in the new world.