In the past, I imagined a scenario where the maturation of AI patent drafting tools would reduce corporate in-house counsel’s reliance on external law firms. For time-sensitive cases, companies might opt to use AI for initial drafting, with outside firms only handling the filing. After all, in-house counsel understands the technology better and this approach would save on the friction costs of external communication.
As more advanced tools become available, in-house teams would become more efficient in areas like technology mining, Freedom to Operate (FTO) analysis, and patent intelligence. They could tackle tasks that were previously unmanageable or too expensive to outsource, thereby increasing their visibility and influence within the company.
Although these tools aren’t exclusively for patent experts, R&D personnel typically lack the incentive to use them, leaving them in the domain of in-house patent professionals.
However, a different trend has emerged in recent years that I didn’t anticipate. Some AI patent tools are targeting a much broader market — not patent professionals, but R&D departments. These systems are designed for engineers, enabling them to identify technological trends and breakthrough points from patent data. They can then formulate technical proposals, draft invention disclosures, and even generate patent applications directly. Features like prior art searches and infringement analysis are also being continuously optimized with user-friendly interfaces.
This development is intriguing. Guided by this software, R&D personnel could naturally become “users” of patent information. Could a day come when the technical reports produced by R&D teams “incidentally” include the analysis that patent professionals are currently responsible for? This is a future I had not previously considered.
Nevertheless, even if this were to happen, it would be far in the future. Patents are inherently complex; even I get confused sometimes. How could a layperson grasp them so easily?
No matter how powerful AI becomes, there is a significant difference between appearing to understand something and truly understanding it. They are two completely different things.