AmberScope is a new patent search engine that allows you to easily search for similar patents to a known patent via its network of citation connections. But given the many already excellent patent search engines out there, why did we decide to develop a new, and very different way to search patents?
1) Because there are many undeserving patents out there
As discussed in an earlier post, there is evidence that suggests that many patents are being granted which may not deserve to be granted:
- Only 54% of litigated US patents are valid: Allison and Lemley, 1998
- Only 22% of US patents have all claims unchanged after re-examination: USPTO, 2012
- 40% of oppositions to granted European patents lead to full or partial loss of claims: EPO, 2012
- Only 47% of litigated Australian patent claims survive invalidity proceedings: Wetherall and Jensen, 2005
- Many patent attorneys believe that non-litigated patents are also of doubtful validity: "the percentage of invalid patents -probably more than 60%" - US patent litigator, 2012
Yet these undeserving patents can have a distortionary effect on the markets they are filed in. Competitors may be discouraged from releasing products and services they have the right to release, or need to commit time and money to file undeserving patent assertions.
The awareness of these undeserving patents is now starting to become a public issue, recognised by the general public as well by the patent profession. In turn, this is starting to lead to political pressure to change the patent system. So it may be in the interest of all of us in the patent profession to improve the quality of granted patents, so that patent laws are not changed in a detrimental fashion as a knee-jerk reaction to undeserving patents.
2) Because searching long lists of patents can be boring and unproductive - and risky
Many likely readers of this blog would have had to search long lists of patent for relevant art. We all know that it can be a boring and unproductive use of our time. So we try to reduce the length of the lists by making certain assumptions about keywords and IPC codes, but these can create the risks that these assumptions lead to relevant prior art patents being missed.
To minimise these risks, patent examiners can be very careful about patent searching, one of the key parts of their job. This can lead to patent examination taking longer than it otherwise might, creating a backlog of patent examinations that can run to 5 years or more in some offices. This in turn provides greater uncertainty to both patent applicants and in particularly their competitors.
3) Because searching by keywords and patent classification codes can miss relevant patents
Whether intentional or not, different patent applicants (and their patent attorneys) can use different keywords for the same technical concepts. As but one example, a cardboard box can be also be described as a container, carrier, carton, receptacle, package, packaging, and there are probably terms I have missed. This can be repeated for almost every technical concept, and means that even the most diligent of searchers can miss relevant patents, even allowing for the dictionaries of synonyms now available in some search engines.
The same applies to patent classification codes such as IPC or USPTO classifications. Because there is always a degree of subjectivity in the allocation of these codes by patent examiners, different patent examiners (for examples in different patent offices) can allocate different IPC codes to very similar inventions. Further errors can result when patent codes are automatically converted from say the USPTO patent classification system to the IPC system.
So the current patent searching system can be boring, unproductive, and still miss relevant patents for a variety of reasons. This can lead to undeserving patents being granted and long examination backlogs, which can have a variety of economic and even political consequences. Clearly improved patent searching processes are required, and AmberScope is our contribution to improving patent searching.