If AI starts generating more inventions, the central question is not whether a machine can sign an inventor declaration. It is how the patent system should evolve to best serve its economic purpose.
That’s an interesting, incisive and well thought out article with a number of considered approaches to an issue that doesn’t get much attention.
Unfortunately, however, I fear the answer is none of the above.
I have thought about this a lot. I’m a big proponent of AI and we have used it a lot to develop systems that make us very efficient. I have worked with very bright people patenting genuinely clever AI technology that enables some of the advances. I’m a great believer in the value of the patent system in general to small and large businesses and have been involved in obtaining grant of about 5000 patents in my career. I enjoy opposing invalid patents and am not a fan of bad patents becoming a burden and have been involved in knocking out or limiting hundreds of patents.
And my conclusion, so far, from all of that direct and positive experience and genuinely thinking long and hard about it and chewing the fat with my similarly experienced colleagues who like to find solutions to tricky problems is… er, we’re doomed.
I just can’t think of any solution that doesn’t inevitably have unintended consequences either in principle or in the way it will undoubtedly be (a) imperfectly implemented and (b) gamed (possibly even by cynical people like me using AI to advantage for their clients). It makes it an intractable problem of whackamole.
I think we’re not far apart in that assessment.
The subtle difference is the most considered response I have come up with is to make the most of using bespoke AI tools to help me and my colleagues get what we need to get our clients patents faster and more effectively or oppose more effectively and enjoy the ride until the intrusion of AI in all other areas of life including the general erosion of democracies bring about a human catastrophe that makes a little dysfunction in the IP system seem somewhat less important.
Have a beer. Ride a motorcycle. Get a few more patents granted while you can. Anything else is likely to be futile and lead to frustration.
Hope my somewhat flippant or fatalistic response to a genuine issue is taken in the spirit intended.
But if you do persuade me you have a practical viable solution I will (a) buy you a beer and (b) sincerely lend my full support and try to mobilise my many friends in the industry to do likewise.
Your line captures the difficulty very well: "I just can’t think of any solution that doesn’t inevitably have unintended consequences either in principle or in the way it will undoubtedly be (a) imperfectly implemented and (b) gamed."
Part of me feels exactly the same. Even if we can imagine a first-best solution in theory, implementation will almost certainly be second- or third-best (your point a) and then subject to strategic behavior (your point b). We have called this the Realpolitik of the patent system in our paper with Jaffe and Wasserman.
But I will still side, cautiously, with the optimists. I think we may find solutions that are imperfect, administratively messy, and sometimes gamed. But these solutions will still be better than doing nothing. The standard should not be perfection; it should be whether we can nudge the system in a direction that better promotes the progress of science and the useful arts...
And this may still be an easier problem than fixing the general erosion of democracies 😕
Even if we don't find a solution, I will gladly have a beer if you ever come to Lausanne!
That’s an interesting, incisive and well thought out article with a number of considered approaches to an issue that doesn’t get much attention.
Unfortunately, however, I fear the answer is none of the above.
I have thought about this a lot. I’m a big proponent of AI and we have used it a lot to develop systems that make us very efficient. I have worked with very bright people patenting genuinely clever AI technology that enables some of the advances. I’m a great believer in the value of the patent system in general to small and large businesses and have been involved in obtaining grant of about 5000 patents in my career. I enjoy opposing invalid patents and am not a fan of bad patents becoming a burden and have been involved in knocking out or limiting hundreds of patents.
And my conclusion, so far, from all of that direct and positive experience and genuinely thinking long and hard about it and chewing the fat with my similarly experienced colleagues who like to find solutions to tricky problems is… er, we’re doomed.
I just can’t think of any solution that doesn’t inevitably have unintended consequences either in principle or in the way it will undoubtedly be (a) imperfectly implemented and (b) gamed (possibly even by cynical people like me using AI to advantage for their clients). It makes it an intractable problem of whackamole.
I think we’re not far apart in that assessment.
The subtle difference is the most considered response I have come up with is to make the most of using bespoke AI tools to help me and my colleagues get what we need to get our clients patents faster and more effectively or oppose more effectively and enjoy the ride until the intrusion of AI in all other areas of life including the general erosion of democracies bring about a human catastrophe that makes a little dysfunction in the IP system seem somewhat less important.
Have a beer. Ride a motorcycle. Get a few more patents granted while you can. Anything else is likely to be futile and lead to frustration.
Hope my somewhat flippant or fatalistic response to a genuine issue is taken in the spirit intended.
But if you do persuade me you have a practical viable solution I will (a) buy you a beer and (b) sincerely lend my full support and try to mobilise my many friends in the industry to do likewise.
Thanks for sharing your thoughts, Ikya!
Your line captures the difficulty very well: "I just can’t think of any solution that doesn’t inevitably have unintended consequences either in principle or in the way it will undoubtedly be (a) imperfectly implemented and (b) gamed."
Part of me feels exactly the same. Even if we can imagine a first-best solution in theory, implementation will almost certainly be second- or third-best (your point a) and then subject to strategic behavior (your point b). We have called this the Realpolitik of the patent system in our paper with Jaffe and Wasserman.
But I will still side, cautiously, with the optimists. I think we may find solutions that are imperfect, administratively messy, and sometimes gamed. But these solutions will still be better than doing nothing. The standard should not be perfection; it should be whether we can nudge the system in a direction that better promotes the progress of science and the useful arts...
And this may still be an easier problem than fixing the general erosion of democracies 😕
Even if we don't find a solution, I will gladly have a beer if you ever come to Lausanne!