Article reused with permission from WIPO.

By C. Andrew Keisner, Consultant, Julio Raffo and Sacha Wunsch-Vincent, Economics and Statistics Division, WIPO.

Robotics and artificial intelligence (AI) have huge potential to transform many aspects of our lives. Recent trials with humanoid robots in supermarkets, schools, hospitals and retirement homes in Europe, the United States and Japan give a sense of this. Hollywood movies like Ex_Machina and Her have also attracted public attention, raising questions about the potential superiority of robots and AI.

But how is innovation in robotics taking place, and what is the role of intellectual property (IP) in that process?

Robotics, the field of technology that drives the development of robots, has played a role for decades in automotive factories, construction sites, schools, hospitals and private homes. But more recently, newer fields of research, including AI and sensing, have combined with robotics to produce advanced autonomous robots with many more potential applications.

What exactly is a robot?

Generally speaking, a robot has the ability to interpret its environment and adjust its actions to achieve a goal. The first modern-day robots were invented for programmed automation to speed up industrial manufacturing processes. But robots have now become fully autonomous systems that can operate and make “decisions” without human interaction.

By 1970 robotic manufacturing was widespread across the automotive industry in the United States and Japan, and by the late 1980s Japan had become the world leader in the manufacture and use of industrial robots. Since the 1980s, when the use of industrial robots in automotive and other industrial production lines became commonplace in the United States and Japan, robotic mechanical parts have become ever more sophisticated and increasingly autonomous. New materials and cutting-edge breakthroughs in artificial intelligence, mechatronics, navigation, sensing, object recognition and information processing have transformed robotics into a multidisciplinary field. Robots are becoming increasingly data-driven and linked over intelligent networks, such as those being developed for self-driving vehicles and drones.

The impact of robots

Robots are already having a significant impact on manufacturing processes in the automotive and electronics sectors. They are also increasingly being used in agriculture, mining, transport, space and sea exploration, unmanned surveillance, health, education and many other fields.

Robots can increase labor productivity, reduce production costs and improve product quality, and in the service sector have spawned entirely new business models. Robots also contribute to human welfare by eliminating the need to do strenuous or dangerous work, by supporting an ageing population and by making sustainable transport a reality.

The industrial robot market, including the cost of software, peripherals and systems engineering, is set to increase to around USD 33 billion by 2017. Asia (in particular, China, the Republic of Korea and Japan) is the world leader in global robotic sales, followed by Europe and North America. The economic gains derived from using robots are directly linked to replacing part of the workforce. While the productivity gains generated by robots help keep firms competitive and create higher-wage jobs in some countries, the overall impact of robots on employment remains uncertain, and quantifying the economic benefits of their use remains a challenge.

The robotics innovation system

Robotics innovation is concentrated in a small number of countries and clusters typically centered around leading universities. Examples include Boston (United States), the Île-de-France (France), Odense (Denmark), Zurich (Switzerland), Bucheon (Republic of Korea), Osaka (Japan) and Shanghai (China). These clusters thrive on the interface between public and private research, with firms commercializing innovations developed partly through long-term research in universities and other public research organizations.

Most robotics-related innovation and company startups are found in high-income countries, with the exception of China, which hosts some of the fastest-growing robotics companies such as DJI (a drone company), Siasun and Estun.

The robotics innovation ecosystem is highly dynamic, research-intensive and collaborative and is becoming increasingly complex. It involves an expanding network of specialists, research institutions and technology-intensive firms, large and small, and brings together know-how from a diverse range of fields to deliver ground-breaking inventions built on the latest developments in materials science, motive power, control systems, sensing and computing.

The collaborative nature of robotics innovation is due in part to the extremely complex challenges presented. Often, companies simply do not have all the required expertise in-house and have to look outside to secure it, for example by establishing joint development agreements with specialized robotics companies.

Industrial robotics is capital-intensive. Research can take years to bear fruit, but university spin-off companies formed around different breakthroughs are driving the sector’s evolution.

Larger, established companies like ABB (Switzerland), Kawasaki Heavy Industries, Yaskawa and Fanuc (Japan) and KUKA (Germany) are also very active in robotics R&D. Large companies active in defense, aerospace and security have also gained expertise in robotics, along with consumer electronics firms like Samsung (Republic of Korea) and Dyson (United Kingdom).

And as robotics becomes more reliant on connectivity and ICT networks, firms like Amazon, Google, Facebook, Infosys, Alibaba and Foxconn are also joining the fray. Many companies in many sectors are beginning to recognize the benefits of robotics, which are increasingly at the heart of business strategies.

Robotics innovation and intellectual property

As more players enter the robotics ecosystem and as innovation focuses on more advanced robotics, companies are increasingly turning to the tools of the IP system to safeguard their interests. Compared to the standard industrial robot innovation of the past, robotics innovation today involves more actors, more technology fields and many more patent filings. Offensive and defensive IP strategies are becoming more commonplace.

Patent protection can be particularly important in this field, given the capital-intensive nature of R&D prior to commercialization and the need for regulatory approval. It allows companies to recoup their investment and helps them secure a competitive commercial advantage. It is particularly useful in protecting inventions that can easily be reverse engineered.

A solid patent portfolio makes it possible to license and cross-license technologies, and thereby strengthen business relations, generate new revenue streams and, in some cases, help avoid litigation. It can also help small firms attract much-needed investment. Robotics patenting surged in the 1980s, when widespread factory automation resulted in a quadrupling of patent applications. Patent applications surged again in the mid-2000s as more advanced robotics came on stream.

Automotive and electronics companies remain the largest filers of robotics patents but new actors are emerging. University–industry collaboration remains strong as the stock of patents held by universities and public research organizations offers significant opportunities for commercialization. Extensive cross-fertilization of research remains a feature of the robotics innovation ecosystem, but there is evidence that patenting is supporting the specialization of firms. Such specialization is important in driving the sector’s continued evolution.

Many robotics companies are using patent documents to find out about the latest technological developments, to gain insights about competitors’ strategies and to monitor whether competitors’ patent claims need to be challenged.

Trade secrets and robotics

The technological complexity of robotics systems means that trade secrets are often the first option for companies seeking to protect their innovations. This makes sense for a number of reasons:

  • Few people have the expertise to reverse engineer these complex systems.
  • Highly expensive robots are very difficult to get hold of, making reverse engineering practically impossible.
  • Many smaller companies want to avoid the cost of filing patent applications.
  • Historically, companies seeking to patent protect their technological advances – especially those decades away from use in market-ready products – spent a lot on patenting but reaped little reward because by the time they were commercialized many of the patents had expired.
    In the robotics sector employee mobility is high, so many companies apply restrictive covenants when employees move to competitors.
  • Uncertainties surrounding the patentability of software in different jurisdictions could further tilt the balance in favor of trade secrets.
  • Next to patents, industrial designs that protect a robot’s appearance – its shape and form – also play an important role in improving the marketability of products and helping firms appropriate the returns on their R&D investments.

Being first to market, a strong after-sales service, reputation and brand have all been critical to the success of past robotics innovation, and remain so today, especially as the industry moves towards developing applications with direct consumer contact. Strong brands are particularly important when selling directly to end-users. That is why most robotics companies trademark their company names and those of their robots.

Copyright, the traditional means of protecting software code, is also relevant to robotics. Under the 1996 WIPO Copyright Treaty, circumventing a technological protection measure to access copyrightable computer code is not permitted. This is of particular relevance to the robotics industry because most companies employ electronic barriers to restrict access to their computer code.

Robotics platforms and the coexistence of IP and open source

Today’s robotics innovation ecosystem is made up of a combination of open and competitive (proprietary) approaches to IP management.

At the pre-commercial stage, a lot of innovation is built around collaborative, open platforms like the Robot Operation System (ROS). These platforms invite third parties to use and/or improve existing content under open licensing arrangements (e.g. Creative Commons, GNU General Public License or a free software license). This allows for rapid prototyping and experimentation. Collaborative platforms enable users to share substantial up-front investment costs, avoid duplication of effort and perfect existing approaches. Such platforms exist for both software and hardware development.

But when innovative firms invest in their own R&D efforts, they tend to protect their inventions more vigorously, especially when they are used to differentiate end-products. As the commercial stakes rise, it will be interesting to see if robotics companies change their approach to IP management.

Will robots ever hold rights in their inventions or creative works?

In future, robots are likely to produce new solutions to problems and in so doing to create intangible outputs that could, at least in theory, be perceived as IP. This could raise interesting questions as to the boundaries of the current IP system. Do objects, software code or other assets created autonomously by a robot qualify for IP protection? If so, how? And who would own these IP rights? The producer? The user of the robot? The robot itself?

Some countries, for example Japan and the Republic of Korea, are actually considering extending rights to machines. In New Zealand the law suggests that original works, even if created by software, robots or AI systems, are protectable under New Zealand’s 1994 Copyright Act. However, such works would belong not to the robot or intelligent system, but to the person(s) who created or utilized the robot or intelligent system that created the work. In other jurisdictions, like the United States, it seems unlikely that a work created by a robot would qualify for copyright protection. Already, contradictory rules relating to the protection of robot-generated IP are emerging among nations that are playing a significant role in robotics development.

Autonomous robot creation and the question of who owns IP rights in creations produced by robots will surely be a matter of much future discussion.