Data: the fuel transforming the global economy

By James Nurton, freelance writer

Data has the power to drive innovative products and services, improve society and tackle health and environmental challenges but what does it take to balance extracting value from data with protecting IP and other rights.

What did you have for breakfast this morning? How tall are you? What’s the tire pressure in your car? When was the last time you shopped online? What is your genetic fingerprint? There is data in every aspect of our lives, and in the past few years, our ability to store and analyze data has increased exponentially. That raises complex questions such as: how do you value data, how can that value be extracted, how can it be shared safely and what rights need to be recognized and protected?

These were some of the topics addressed at the fourth session of the WIPO Conversation on IP and Frontier Technologies, which took place in September 2021 (day one and day two available). Frontier technologies include artificial intelligence (AI), Blockchain, the Internet of Things and robotics, and they are estimated to become a USD 3.2 trillion market by 2025.

More than 1,600 people from 130 countries registered to take part in the session. Five hours of discussions ranged over the protection and regulation of data, its role in artificial intelligence and how to balance access and control, with examples of how data is being used in research and business. These ranged from using AI to produce music to monitoring the behavior of bees in beehives (read “Uncanny Valley: charting a new ear of musical creativity” and “Beewise: out-of-the-box thinking to save the world’s bees”).

Introducing the meeting, WIPO Director General Daren Tang said digitalization has been accelerated by the COVID-19 pandemic: there are expected to be 43 billion devices connected to the Internet of Things by 2023, and there are over 1 million new 5G subscribers every day. “If digitalization is the engine of the future economy, then data is its fuel,” he said, noting that 60 percent of the world’s population is now online: “The increasing connectivity and the resulting stream of data is powering the rise of frontier technologies. In this inter-connected world, it is crucial we understand the nature of data and its value.”

The value of data

A striking example of the power of data was given by Dean Jolliffe of the World Bank, who chaired the meeting’s first panel, “Data, beyond AI in a fully interconnected world”. In 1999, Cyclone BOB 06 devastated the Indian state of Odisha, killing nearly 10,000 people. In response, the state disaster management authority was tasked with collecting, evaluating and analyzing weather data. When a similar-sized cyclone hit Odisha in 2013, over 1 million people were evacuated, and thousands of lives saved.

The Odisha case shows, said Mr. Jolliffe, how the value of data – even data that has been collected and tracked for many years – can increase when rigor is applied: “By building up a data system that made data more interoperable and accessible and enabling the findings to be communicated in ways that were timely and understandable, this data became priceless.”

With smart devices and sensors now ubiquitous, data can transform every aspect of business and society. But, said Aruba Khalid, Senior Research Analyst at Dubai Future Foundation in the UAE: “The value of data depends a lot on the insights that it’s used to drive and to some extent on the scale at which those insights are applied.”

Ms. Khalid said that data can be used to streamline operations and improve cost efficiency (for example, aircraft manufacturer Airbus has cut supplier delivery times from a couple of weeks to just a couple of hours by sharing design and engineering data). It can also be used to enable new businesses based on creating tailored products (such as Amazon, Netflix and Facebook) and it can create entirely new sectors and businesses that would otherwise be unimaginable: examples include all the industries that rely on data from satellite information and the emerging area of precision medicine.

Monetary and social value

As these examples show, innovative products often require data from multiple sources to be combined and building walls may prevent this value being realized. Diane Coyle, Bennett Professor of Public Policy at the University of Cambridge in the UK, argued that not only is data sharing necessary but there are also arguments for making certain data freely available. Data has monetary value but also social value; the latter is hard to put a figure on and may not attract private investment. For that reason, governments have for a long time provided data such as national statistics as a public good. “There will be better outcomes that can be achieved by policy interventions. There might be social value that can’t be captured by private companies and individuals,” said Professor Coyle.

Other speakers agreed. Caroline Wanjiru Muchiri, of the Centre for Intellectual Property and Information Technology Law, Strathmore University in Kenya, emphasized the contexts affecting different countries. In some African countries, for example, public service delivery depends on access to data, while organizations such as religious institutions and humanitarian agencies hold large volumes of critical data. In some countries, colonial institutions continue to play a role, and this has consequences including a lack of data about women.

As data has most value when it is accumulated, Jiro Kokuryo, Professor at the Faculty of Policy Management at Keio University in Japan, asked whether Western-style approaches to ownership based on private property rights may be less appropriate than Asian-derived principles of benevolence. Under these theories, data might best be understood as belonging to society as a whole, rather than to individuals, he said: “We would emphasize harmony and respect, instead of individuals’ rights and autonomy.” Or maybe a third way can be found, involving loyalty to whoever provides the original data? “We need to think about what the ethos of the digital economy should be,” said Professor Kokuryo.

Structured and interoperable

The ability to accumulate vast amounts of data (“big data”) is due to the expansion of our digital footprint; it is driven by the interconnectivity of devices and cheaper storage, and it is likely to accelerate. WIPO Director General Tang said that 90 percent of all data in the world has been created in the past two years, and every day the amount of data produced is 2,500 times greater than what is stored in the British Library. But the scale of this data raises its own challenges: how can the relevant data be identified and how do we promote interoperability, ensure fairness and inclusion, and reduce inefficiencies in sharing?

Context is key, said Professor Coyle: “My temperature only provides useful information if I know what the population average is.” Medical data may be very powerful in diagnosing or predicting disease, but only when it is combined with other information. Some data can retain value even after it is used, but some – such as traffic or weather data – can depreciate very quickly.

A granular understanding of data, taking account of the different context, requires the creation of structured data, which can be transferred without losing its value, and procedures to promote interoperability. In his speech, Kung-Chung Liu, Professor of Law at Singapore Management University in Singapore, proposed a unified standard for data formatting to promote trading of data including across borders.

Medical data may be very powerful in diagnosing or predicting disease, but only when it is combined with other information. Some data can retain value even after it is used, but some – such as traffic or weather data – can depreciate very quickly.

The regulatory matrix

Discussions about sharing and trading data prompt difficult questions about how data is defined, categorized and regulated. These involve questions about security, ethics, privacy, ownership and other matters. A complex set of regional and national agreements addressing these. Many of the questions also concern fundamental rights.

For example, the right to privacy is especially relevant for sensitive or personal data. A lot of legislative attention has been paid to protecting privacy ever since the rise of the Internet. An example is the EU General Data Protection Regulation. The right to privacy continues to be a priority for many lawmakers amid concerns over security breaches, respect for confidentiality and loss of control over personal data.

Discussions about sharing and trading data prompt difficult questions about how data is defined, categorized and regulated.

But focusing too much on privacy “increases the danger of missing valuable uses” according to Professor Coyle. Other speakers shared her concerns. Solutions to this dilemma need to be found. They could include ensuring that data is used only for a specific purpose; requiring consent for all uses; establishing principles such as fairness and dignity; or even creating a fiduciary duty of loyalty to the person who provides the data. Whatever the solution, data must be traceable so that integrity can be assured.

As individuals, businesses and governments address these issues it is worth taking a step back to consider the values driving the use of data. As Dafna Feinholz, Chief of Bioethics and the Ethics of Science at UNESCO, said in her speech: “Law is codified ethics, and AI technologies are not value-neutral.”

In November 2021, UNESCO’s General Conference adopted the Recommendation on the Ethics of Artificial Intelligence, following a multi-stakeholder consultation process and input from member states. The Recommendation includes 10 principles to ensure that “the development and use of AI technologies are guided by both sound scientific research as well as ethical analysis and evaluation.” The Recommendaiton also covers data privacy, sharing and governance and promises to provide a template for the continuing data debate.

Where do IP rights fit in?

Most data does not fit neatly into existing regulatory frameworks such as existing intellectual property (IP) systems. As Bret Hrivnak, of The International Association for the Protection of Intellectual Property (AIPPI), said: IP rights protect “creations of the mind” but most data is non-creative. Patents can protect processes for using data or how it is generated, but not data itself, while copyright protection may protect some types of data but generally only where there is structure and originality. “These IP rights are limited – the alternative is trade secrets and contract law,” said Mr. Hrivnak.

Trade secrets provide “a flexible system” that can protect data, said Elisabeth Kasznar Fekete, Senior Partner of Kasznar Leonardos in Brazil, adding that she is seeing “more precise contracts” setting out the ownership and licensing of data, and the balancing of rights. “We need to create integrated systems and models between the owner of the data and those interested in licensing,” she said. “IP has to follow the path of everyday business.”

Data poses complex issues for the IP system. IP can pose both an incentive for the investment in generation of data, for example through database rights, or a hindrance to access to data, due to limited text and data mining provisions.

In the EU, there has been sui generis protection since 1996 to reward the investment made in producing databases, but there must be substantial investment in the collection and compilation of the data that makes up the database. “Could database protection have a role in legal protection of data generated by AI though data mining? Is the processing made by an AI system data creation or data obtaining?” asked Tatiana Eleni Synodinou, Associate Professor in Private and Commercial Law at the University of Cyprus, who concluded: “It’s not clear whether the law protects these kinds of databases.”

Data mining – the process of extracting large sets of data – raises several controversial issues, particularly where the data involved is or may be protected by copyright. Some countries, notably Japan, have provided broad exceptions to copyright for text and data mining to promote AI development. Under US copyright law, however, whether or not data mining is allowed falls under the general fair use exception. In the EU, a text and data mining provision has been implemented that allows mining for research. However, the distinction between commercial and non-commercial uses can lead to uncertainty. As Carlo Scollo Lavizzari, Partner of Lenz Caemmerer in Basel, Switzerland, said: “The future of reading is mining – and the future of mining is content worth mining.”

Conclusion: IP in context

The range of the discussion demonstrated how broad and complex are the issues around data. The final part of the meeting featured statements from stakeholders from around the world on many topics, including how the IP system can support data-driven innovation, text and data mining, the need for transparency and trust and how to strike the right balance between competing interests. It demonstrated the wide impact of these questions, and the real concerns that many people have.

The range of views expressed showed the extent to which data poses complex issues for the IP system. IP can pose both an incentive for the investment in generation of data, for example through database rights, or a hindrance to access to data, due to limited text and data mining provisions.

The Conversation highlighted the important role played by IP alongside other considerations in regulating data. While WIPO’s mandate relates directly to the development of a balanced and effective IP system, all these other parts of the picture need to be kept in mind to ensure a coherent and consistent approach to these questions.

Acknowledgements: Ulrike Till, IP and Frontier Technologies Division, WIPO