Austrian Research and Technology Report 2020

174
Austrian Research and Technology Report 2020
language generation. The same survey is expected to 
include a dedicated module on AI in 2021 that will ask 
questions about the technologies used, the purpose 
of this use and whether the technology applied was 
developed by the company itself or by external pro-
viders. In February 2020, the European Commission 
and the OECD also agreed to harmonise the EU’s AI 
Watch platform and the OECD’s AI Policy Observato-
ry (OECD.AI) in terms of the information used in and 
for them. The first phase of collaboration focused on 
building a database of national AI strategies and pol-
icies. In connection with the coordinated AI plan, the 
EU Member States have committed to developing na-
tional AI strategies to dovetail their AI policies and 
investments. The next phase will concentrate on mak-
ing the reports from AI Watch and other EU publica-
tions available via the AI Policy Observatory, ex-
changing data more comprehensively and working 
more closely on designing improved data collection 
methods.
176
The WIPO Technology Trends
177
series, 
which builds on the expertise that the World Intellec-
tual Property Organization (WIPO) has in analysing 
patent data, published analyses on the global trends 
in AI issues in 2019. This publication is one of the first 
to systematically investigate trends in AI technology. 
It analyses which areas are demonstrating the largest 
amount of innovative AI activities, which companies 
and institutions are leading the way in AI develop-
ment and where the growth markets of the future will 
be. To this end, WIPO has devised a new framework 
for understanding trends in this area, with AI-related 
technologies being divided into groups to reflect 
three dimensions of AI: techniques used in AI (e.g. 
machine learning), functional applications (e.g. speech 
processing and computer vision) and areas of applica-
tion (e.g. telecommunications and logistics). For each 
of these areas, the report provides data and analyses 
that highlight trends, key players, geographical distri-
bution and market activities, including acquisitions 
and legal disputes.
176 See 
https://ec.europa.eu/digital-single-market/en/news/european-commission-and-oecd-collaborate-global-monitor-
ing-and-analysis-artificial-intelligence
177 See World Intellectual Property Organization (WIPO) (2019).
The most important finding from this analysis is 
that AI-related inventions are booming and are shift-
ing from the theory books to the commercial market: 
the ratio of academic papers to concrete inventions 
fell from 8:1 in 2010 to 3:1 in 2016. Since artificial in-
telligence emerged in the 1950s, innovators and re-
searchers have registered nearly 340,000 inventions 
with an AI element. The patenting of AI-related in-
ventions is increasing sharply: over half of the inven-
tions identified have been published since 2013. As 
well as revealing AI techniques and applications, 
AI-related patents are often also connected to a spe-
cific area of application or industry. WIPO’s analysis 
showed that many sectors and industries are con-
ducting research into the commercial use of AI. It 
identified 20 different areas of application, with at 
least one being cited in 62% of all the AI patent data 
collected. These included, in descending order of 
size: telecommunications (cited in 15% of all patent 
documents identified), transportation (15%), life and 
medical sciences (12%), and personal devices, com-
puting and human-computer interaction (11%). Other 
sectors were banking, entertainment, security, indus-
try and manufacturing, agriculture, and networks (in-
cluding social networks, smart cities and the Internet 
of Things). Companies, particularly those from Japan, 
the USA and China, are the dominant forces in pat-
ent activity. In total, there are 26 companies in the 
top 30 applicants for AI-related patents, compared 
with only four universities and public research insti-
tutions. This is the case for most AI techniques, ap-
plications and fields. Of the top 20 companies apply-
ing for AI-related patents, twelve are based in Japan, 
three are from the USA and two are Chinese. Japa-
nese firms in the entertainment electronics industry 
are especially well represented. Despite companies 
dominating the AI field, universities and public re-
search organisations are playing a leading role in 
coming up with inventions in selected areas of AI. 
Chinese organisations make up 17 of the leading 20 

---