Send an email
Send an email
Send an email
Whilst the number of AI filings per year at the EPO has continued to rise, the pace of growth has slowed, with a 24% rate of growth in 2019, compared to 45% in 2017, suggesting that the AI industry is beginning to mature. So, what does this mean for AI’s next phase of growth?
Uptake of Technologies
If we take the total number of AI patent filings at the EPO, the largest technology sectors were Life & Medical Science (15%), Telecommunications (12%) and Physical Sciences (11%). Interestingly, when looking at the change for individual sectors since our previous report in 2021, the manufacturing and industry sector has risen by three places into the top 10 sectors, moving from 4% of total filings at the time of our last report, to 6% of total filings now, which is indicative of AI being applied more in the physical economy.
The trend of increased numbers of AI applications was maintained across all but one category since 2021. However, although only one sector experienced a decrease in publications, a number of other sectors have experienced a reduction in the rate of growth of publications in recent years. The only exceptions to this were Life Sciences, Physical Sciences, and Industry and Manufacturing which were the only three sectors to see a continuation of their growth rates from the last few years.
Interestingly, it is noted that “Industry and Manufacturing” had a relatively low acceptance rate, which may be due to a large number of AI applications relating to planning and scheduling in this field – such aspects are more likely to be deemed “non-technical” by the EPO. Generally, the sectors with above-average allowance rates are those that the EPO has historically deemed to be “technical”, while the sectors with below-average allowance rates are those the EPO has deemed “non-technical”. This reflects the EPO’s requirement for an invention to provide a “technical effect” in order to be patentable.
Looking at types of technology, computer vision has shown a striking increase in AI patent publications over the last five years. Conversely, speech-processing publications have been in decline. This likely suggests that after a period of heavy investment, speech processing has reached a maturity commensurate with the current market opportunities, while the heavy increase in computer vision likely indicates a still significant runway of potential applications.
The Global Picture
Looking more closely at the number of applicants on a country-by-country basis, Chinese applicants at the EPO overtook Japanese applicants in 2020, making it the third largest filer of AI applications at the EPO, behind US and European applicants. However, between 2015 and 2020, per capita, applicants from the Republic of Korea have overtaken of Japan, the United States and Europe to become the largest filers on a per capita basis.
Looking at the success rate, US applicants have consistently had a lower success rate than applicants from any of the other top 5 countries over the last 20 years. This could reflect a larger proportion of “non-technical” patent applications originating from the US. Having said this, in our previous report, we showed that US originating AI applications had a lower allowance rate before the EPO than European originating applications even when normalizing for normalising for different proportions of “technical” and “non-technical” cases. Notably, the difference in allowance rate between European applicants and US applicants is greatest for non-technical cases at 11.5%, followed by mixed cases at 8.2% and technical cases at 7.3%. We see this as evidence of the value of applications being prepared with European requirements in mind and it is not surprising to us to see this value being greatest in more challenging non-technical cases where excluded subject matter is more likely to arise.
Increasing Allowance
As the number of AI patent applications continues to increase in Europe, allowance rate becomes one of the most crucial determining factors for AI’s next phase of growth.
The allowance rate for AI applications has continued to increase over the past ten years. There was a particularly large jump in allowance rate in 2021 to 64% (an increase of 9% on the 2020 allowance rate of 55%). Whilst this jump in AI allowance rate represents a relatively large change compared to 2020, it is notable that there was a drop of 2% in allowance rate in 2020. Nevertheless, the allowance rate for AI applications appears to be consistently below the average for all applications at the EPO. For instance, the overall allowance rate for all applications at the EPO was 69% in 2020. The lower allowance rate for AI related applications likely reflects the additional difficulties associated with protecting computer-implemented inventions at the EPO.
Landscape to Change?
As overall patent filings at the EPO in 2020 fell by 0.7% (likely due to the COVID-19 pandemic), it will be interesting to see how AI filings hold up over this period and 2021. Note that, due to the 18-month delay in applications publishing, we do not yet have the data for the whole of 2021 available for analysis.
Similarly, the trend of increasing numbers of AI applications was maintained across all but one category. In particular, having enjoyed remarkable growth since 2000 to 2020, Transportation was the only sector to see a notable decrease in the number of publications in 2021. It is too soon to say whether the decrease in Transportation AI publications sets a new trend (perhaps as a result of increasing difficulty of progress in autonomous vehicles), or if the pace of publications will again pick-up in 2022.
Another notable landmark is the EPO Guidelines for Examination which are updated on annual basis. After the major update to the Guidelines on AI filing trends, the Guidelines were updated in 2022 with the inclusion of a number of examples where technical character can arise. By drawing analogies with these examples, it can be possible to convince EPO Examiners of technical character, and therefore to obtain grant for more borderline cases. The inclusion of these further examples is therefore a welcome addition, which provides additional guidance on what may be considered technical for AI applications. We are hopeful that this improved clarity will help applicants more effectively tailor their applications to focus on technical advantages to improve their prospects of achieving grant at the EPO. It will be interesting to see if this improved clarity on examination requirements improves allowance rate for AI applications going forwards.
Closing Thoughts
The statistics paint a picture of AI as a rapidly maturing technology. After an initial boom in filings over the last few years, the filing figures are starting to show a slow-down in growth. It remains to be seen if this in caused by the market settling down after the initial AI gold rush, or whether this reflects other challenges facing all businesses, such as increased economic uncertainty. However, the continued growth along with a wave of exciting developments mean that the market is far from another AI winter. Further supporting this is the fact that new countries are emerging as frontrunners when it comes to patent applications, with the Republic of Korea coming out top per capita
Early AI success stories, such as speech processing, have seen a slow-down in filings in recent years, with a move to new technology areas such as computer vision. This suggests a changing focus within AI R&D. In addition, recent increases in filings in the Industry & Manufacturing sector suggest a move towards AI being applied more widely in the physical economy.
With increasing allowance rates at the EPO, and increasing clarity from the EPO regarding the tests for patentability, the AI sector is increasingly becoming an important source of patentable innovation.
