Claims are a movable target

My clients are using various AIs to write up their disclosures of their inventive products. Their AIs then write them up as patent applications with a full set of claims at the end. My clients then work hard at these claims, making sure they cover as many versions of their invention as possible.

As I discussed in my article “WHY I CAN’T FILE YOUR AI GENERATED PATENT APPLICATION”, we then reorganize and rewrite these texts into proper patent applications, by focusing on the inventive aspects of the product, and rewriting the claims to match. Our clients then work hard on this new set of claims, possibly with the help of their AIs.

This back and forth about the claims is useful, but only up to a point.

As my friend and late colleague, Larry Rublin z”l, pointed out many years ago, claims change during prosecution but the specification, where we describe what the invention is, how to build it and what happens when you operate it, is fixed. Once it is written, no new matter may be added to the description. Only errors can be fixed. Thus, when writing the specification of the patent application, one should expend most of your energy on the description.

What do I mean by “claims change”?

Claims are amended during prosecution — the back-and-forth between your patent attorney and the patent office while your application is examined. The traditional view of amendments is that they narrow the scope of the invention. Thus, if you imagine that your invention is a pie slice, each amendment is equivalent to a bite taken from the pie slice, thereby reducing the breadth of your invention.

This image comes from the traditional thought that amendments usually are made by adding limitations (additional restrictions added to the claim) from the dependent claims (claiming details of one or more elements in the main claim) and these claims, by definition, narrow the scope of the invention.

Most inventors picture prosecution as a slow erosion of that slice. The Examiner finds a piece of "prior art" (an earlier patent or publication that overlaps your slice), you amend the claim to dodge it, and the slice gets a little smaller. A few rounds of this and the slice is a sliver. The inventor ends up wondering what is left worth protecting.

However, when the issue is one of a misreading or misunderstanding of the language of the claims, the dependent claims don’t help. Instead, I need to rewrite the claims to find better wording for what I define as the invention. The image I have is that the claims move, away from the cited prior art (which, in any case, wasn’t relevant since the Examiner’s understanding of the language was wrong). If we keep with the image of the pie slice, the slice keeps its shape but moves away, to a different part of the world.

Imagine instead that the pie slice sits inside the world of the "specification", the document you filed with the patent office. The world contains everything you disclosed: the background, the problem, the solution, the variations, the drawings. The slice is just the region of the world you initially chose to claim. There is more inside the world than what you initially put inside the slice.

Now when the examiner finds prior art that overlaps the slice, there is a better response. Instead of biting into the slice, you slide the slice within the world, into a region the prior art does not touch. You haven't eaten the slice. You've moved it.

How can you do that? Because there is more information in the world than what you put in the original slice. When you drafted the application, you described variations and details that did not all make it into the first set of claims. Those details are sitting there, in the spec, waiting. When the examiner pushes back, we reach into the description and pull those details into the claim. The claim moves to a place defined by features the prior art does not have.

There is a catch, of course. You can move the slice within the world, but not outside it. Anything you did not describe when you filed is not available later. The patent office will not let you claim what your application did not disclose. So the size of the world — the breadth of the description — is the size of your room to maneuver.

This is why Larry used to argue that the description is the more important part of the application. The claims can change; the description cannot. The description is the territory. The claims are just where you happen to be standing in it.

I think Larry was right. Claims are a movable target and the spec is the ground they move on.

Therefore, the client’s attention is essential in preparing the description. The client usually provides an initial discussion of the problem and the solution, from which the patent attorneys may both expand the scope, often adding a variety of terms, and may delve into the implementation details and possibly manufacturing considerations.

The client’s attention is needed in reviewing the text to ensure that all the details are correct (both at the general level and in the detailed implementation discussion).

Many patent attorneys write their patent applications by writing their claims first so that they fully understand what is patentably inventive about the invention. Only then do they write the patent application and they make sure not to write anything that isn’t needed to support their claims. I strongly disagree with this method, largely for the reasons described above. Claims change and the text needs to be written to support those changes.

Investors often ask why the examiner found something which was not found in a pre-filing search. The answer is that their claims moved in response to the results of that pre-filing search. As a result, the claims that the investors are seeing are not the definition of the invention that we used when we did the pre-filing search.

Similarly, the applicant’s understanding of what was so brilliant may change from the initial design to the final product. To handle this, we may either amend the current set of claims, or we may file continuation applications (i.e. applications which have the same filing date as the current application but are filed later) to recite this new understanding of the invention.

We are not required to amend the claims. We could always just argue that the original set of claims were patentable as they were. However, when we pay attention to how others hear the language we use in the claims, to how they are understanding the invention and/or to how we change our understanding of what is inventive, we may find better ways to amend the claims.

Remember - claims are a movable target and the spec is the ground they move on.